Surface-Enhanced Raman Spectroscopy Integrated Centrifugal Microfluidics Platform ·...

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: May 12, 2021

Surface-Enhanced Raman Spectroscopy Integrated Centrifugal Microfluidics Platform

Durucan, Onur

Publication date:2018

Document VersionPublisher's PDF, also known as Version of record

Link back to DTU Orbit

Citation (APA):Durucan, O. (2018). Surface-Enhanced Raman Spectroscopy Integrated Centrifugal Microfluidics Platform. DTUNanotech.

https://orbit.dtu.dk/en/publications/b5784815-e07e-4bd2-9283-461c87d02b87


Onur DurucanPhD Thesis January 2018

SURFACE-ENHANCED RAMAN SPECTROSCOPY INTEGRATED

CENTRIFUGAL MICROFLUIDICS PLATFORM

A Thesis

Presented to

The Academic Faculty

by

Onur Durucan

In Partial Fulfilment

of the Requirements for the

Ph.D. Degree in the

Department of Micro- and Nanotechnology

Technical University of Denmark

January 2018

SURFACE-ENHANCED RAMAN SPECTROSCOPY INTEGRATED

CENTRIFUGAL MICROFLUIDICS PLATFORM

Main Supervisor:

Professor Anja Boisen



Co-supervisor:

Associate Professor Tomas Rindzevicius



Co-supervisor:

Dr. Marco Matteucci

Production Manager

NIL Technology ApS

Co-supervisor:

Senior Researcher Michael Stenbæk Schmidt



PREFACE

In this thesis, the summary of research activities conducted throughout the PhD

employment from 15th of January 2015 at the Nanoprobes research group (Department of

Micro- and Nanotechnology, Technical University of Denmark) is presented. The PhD

project was part of the HERMES-High Exponential Rise in Miniaturized cantilever-like

Sensing project founded by European Research Council (ERC), grant no. 320535.

ACKNOWLEDGEMENTS

First of all, I would like to thank my supervisors Professor Anja Boisen,

Associate Professor Tomas Rindzevicius, Senior Researcher Michael Stenbæk Schmidt

and Dr. Marco Matteucci for their continuous support and guidance throughout my PhD

study. I have really enjoyed the process of my PhD study. There was always somebody

in my supervisor team who would suggest ideas and mentor me to overcome toughest

obstacles that came across the project. I am very grateful to Associate Professor Tomas

Rindzevicius and Senior Researcher Michael Stenbæk Schmidt for countless times of

scientific discussions which provided insights and greatly assisted the research. I have

learned a lot from Professor Anja Boisen, especially about the effective communication,

educational development and personal problem-solving skills. Also, I appreciate my

former supervisor Dr. Robert Burger, who is now a part of a start-up company, for

introducing me to the centrifugal microfluidics platform. His supervision and shared

knowledge during the first year have greatly facilitated the experimental studies and

provided me deep conceptual understanding of microfluidics.

I would like to express my gratitude to the Nanoprobes group which I was part

of during the PhD study for a collaborative, friendly and positive research environment.

On this account, I greatly appreciate my research collaborators and colleagues, especially

Postdoc. Kaiyu Wu, Dr. Kuldeep Sanger, Dr. Anil Haraksingh Thilsted, Lidia Morelli,

Marlitt Viehrig. Also, Rokon Uddin, Lukas Vaut, Postdoc. Oleksii Ilchenko, Zarmeena

Abid for numerous scientific discussions, expertise and support throughout the project.

Lastly, I want to thank my family and Yana for their great support. They have

always kept me positive and motivated and encouraged me to achieve my goals.

Onur Durucan

January 2018

ABSTRACT

This PhD thesis demonstrates (i) centrifugal microfluidics disc platform

integrated with Au capped nanopillar (NP) substrates for surface-enhanced Raman

spectroscopy (SERS) based sensing, and (ii) novel sample analysis concepts achieved by

synergistical combination of sensing techniques and miniaturized fluid handling devices

for facile quantitative and qualitative analysis of real-life sample matrices.

A nanofiltration approach based on a wicking (wetting) phenomenon of fluids

on nanostructured surfaces was introduced. The method provides purification of complex

suspensions by passing it through dense array of NP structures. Furthermore, the wicking

assisted nanofiltration procedure was accomplished in centrifugal microfluidics platform

and as a result additional sample purification was achieved through the centrifugation

process. In this way, the Au coated NP substrate was utilized as nanofilter and SERS

active surface at the same time.

To evaluate the efficiency of the nanofiltration technique, a well-known

application case, detection of toxic melamine molecules in milk, was selected. According

to the statistical SERS map analysis conducted on the purified sample region of the NP

surface (2×4 mm2), the spectral response for melamine was 14 times higher as compared

to the immersed and non-purified part of the chip. The quantitative study of melamine

content with respect to the averaged peak intensity at 687 cm-1 was in accordance with

characteristic Langmuir adsorption curve behavior and the detection limit was estimated

to be 10 parts per million (ppm).

The wicking based fluid sample transport was further investigated for

simultaneous detection of multiple components in human urine solutions. Using the

surface sensitivity of the SERS technique, affinity based chromatographic separation of

urine compounds such as creatinine, uric acid and urea was achieved. Additionally, a

unique and novel quantification procedure based on spectral profiles was presented.

Lastly, an alternative sensing approach was carried out using Au coated NP like

structures on fused silica. By combining the advantages of electrochemical (EC) systems

for quantitative analysis and the molecular specificity of SERS, a proof-of-concept study

on detection of paracetamol in phosphate-buffered saline (PBS) was performed.

RESUMÉ

I denne ph.d.afhandling demonstreres: (i) En såkaldt ‘centrifugal microfluid

disc’ platform med integreret sensor substrat (‘Au capped nanopillar substrate’) til

surface-enhanced Raman spektroskopi (SERS). Og (ii) et nyt analysekoncept hvor

sensor-teknologi og en teknik til håndtering af væske kombineres til en enkel metode til

kvantitativ og kvalitativ analyse af komplicerede prøver.

Der introduceres en metode til nanofiltrering baseret på nanostrukturerede

overfladers evne til at opsuge væske. Med denne metode kan komplekse opløsninger

oprenses ved at passere et tæt array af nanostrukturer (‘Au capped nanopillars’). Denne

metode for nanofiltrering er integreret på førnævnte ‘centrifugal microfluidic platform’

og hermed opnås yderligere oprensning af prøven gennem centrifugeringsprocessen. På

denne måde bruges det integrerede sensor substrat (‘Au capped nanopillar substrate’)

både som filter i oprensningen af prøven og som SERS substrat til detektion, på samme

tid.

For at evaluere effektiviteten af metoden til nanofiltrering, er der valgt en

velkendt anvendelse, nemlig detektion af giftig melamin i mælk. Ifølge SERS analysen

var responsen for melamin på den del af substratet hvor prøven havde passeret

nanostrukturen 14 gange højere end på den del der havde været nedsænket direkte i

prøven. Den kvantitative analyse af mængden af melamin i forhold til den gennemsnitlige

peak intensitet på 687 cm-1, var i overensstemmelse med den karakteristiske Langmuir

adsorption og detektionsgrænsen er estimeret til 10 parts per million (ppm).

Metoden til nanofiltrering er undersøgt yderligere til samtidig detektion af flere

forskellige stoffer i en opløsning af menneskelig urin. Ved brug af SERS teknikkens

overfladesensitivitet opnås affinitetsbaseret kromatografi af stoffer i urin såsom

creatinine, urinsyre og urea. Ydermere, præsenteres der en unik, ny procedure til

kvantificering baseret på stoffernes profiler i spektrene.

Afslutningsvist er de samme nanostrukturer (‘Au coated nanopillars’) på glas,

brugt som en alternativ sensor. Ved at kombinere fordelene ved elektrokemiske systemer

til kvantitativ analyse og den molekylære specificitet fra SERS, er der udført et ‘proof-

of-concept’ studie på detektion af paracetamol i ‘phosphate-buffered saline (PBS)’.

TABLE OF CONTENTS

1 Introduction .................................................................................................................. 1

1.1 AIM OF THE PHD PROJECT ........................................................................................ 3

1.2 STRUCTURE OF THE THESIS ...................................................................................... 4

2 Raman Effect ................................................................................................................ 9

2.1 MOLECULAR VIBRATION MODES ........................................................................... 10

2.2 POLARIZABILITY OF MOLECULES ........................................................................... 13

2.3 RAYLEIGH AND RAMAN SCATTERING .................................................................... 16

2.4 RAMAN SPECTROSCOPY ......................................................................................... 19

3 Surface-Enhanced Raman Scattering ...................................................................... 23

3.1 EM WAVES IN CONDUCTING ENVIRONMENT ......................................................... 24

3.2 SCATTERING ON SPHERICAL PARTICLES ................................................................. 28

3.3 LOCALIZED SURFACE PLASMONS ........................................................................... 31

3.4 EM ENHANCEMENT MECHANISM ........................................................................... 37

4 Plasmonic NP Structures ........................................................................................... 43

4.1 FABRICATION OF METAL COATED SI NP STRUCTURES .......................................... 44

4.2 LEANING EFFECT OF NP STRUCTURES ................................................................... 46

4.3 PLASMONIC PROPERTIES OF NP SUBSTRATES ........................................................ 48

4.4 SENSING APPLICATIONS USING SERS SUBSTRATES ............................................... 50

4.5 ELECTROCHEMICAL SENSING ................................................................................. 52

5 Centrifugal Microfluidics .......................................................................................... 57

5.1 INERTIAL FORCES ................................................................................................... 58

5.2 PNEUMATIC PUMPING............................................................................................. 61

5.3 CAPILLARY VALVING ............................................................................................. 63

5.4 DESIGN AND FABRICATION OF MICROFLUIDIC DISCS ............................................. 65

6 Nanopillar Filters for SERS ...................................................................................... 69

6.1 INTRODUCTION ....................................................................................................... 70

6.2 MICROFLUIDIC PROCEDURE ................................................................................... 72

6.3 WICKING BASED NANOFILTRATION ........................................................................ 74

6.4 QUANTITATIVE ASSESSMENT ................................................................................. 75

6.5 CONCLUSION .......................................................................................................... 77

6.6 SUPPORTING INFORMATION .................................................................................... 78

6.6.1 Experimental Methods ................................................................................... 78

6.6.2 Optimization of Microfluidics Design ........................................................... 82

6.6.3 Characterization of Filtration Technique ....................................................... 84

6.7 ACKNOWLEDGEMENT ............................................................................................. 88

7 NP Structures for SERS Chromatography .............................................................. 91

7.1 INTRODUCTION ....................................................................................................... 92

7.2 SAMPLE HANDLING STEPS ..................................................................................... 94

7.3 SERS CHROMATOGRAPHY PRINCIPLE .................................................................... 96

7.4 SEPARATION EFFICIENCY AND CONCENTRATION DEPENDENCE ............................. 99

7.5 CONCLUSION ........................................................................................................ 103

7.6 SUPPORTING INFORMATION .................................................................................. 104

7.6.1 Experimental Methods ................................................................................. 104

7.6.2 Optimization of Microfluidics System ......................................................... 109

7.6.3 Data Treatment and MCR Analysis ............................................................. 113

7.7 ACKNOWLEDGEMENT ........................................................................................... 116

8 Dual-Functional EC and SERS Sensing ................................................................. 121

8.1 INTRODUCTION ..................................................................................................... 122

8.2 EXPERIMENTAL SECTION ...................................................................................... 124

8.3 RESULTS AND DISCUSSION ................................................................................... 127

8.3.1 Fabrication Process ...................................................................................... 127

8.3.2 SERS Characterization ................................................................................. 128

8.3.3 EC Characterization ..................................................................................... 131

8.3.4 Dual Sensing-Detection and Quantification of Paracetamol ....................... 134

8.4 CONCLUSION ........................................................................................................ 136

8.5 ACKNOWLEDGEMENT ........................................................................................... 137

9 Concluding Remarks ............................................................................................... 141

LIST OF FIGURES

FIGURE 2.1 SCHEMATIC ILLUSTRATION OF A SIMPLIFIED DIATOMIC MOLECULE ............... 10

FIGURE 2.2 VARIOUS “NORMAL” VIBRATIONAL MODES OF THE WATER MOLECULE ......... 12

FIGURE 2.3 POLARIZABILITY IN MOLECULAR SCALE ........................................................ 14

FIGURE 2.4 SCHEMATIC ILLUSTRATION OF THE SCATTERING PROCESS ............................. 17

FIGURE 2.5 RAMAN SPECTRUM OF MELAMINE RECORDED ................................................ 18

FIGURE 3.1 DRUDE MODEL FOR ELECTRON TRANSFER IN CONDUCTING MATERIALS ......... 25

FIGURE 3.2 THE SCHEMATICS OF GENERAL MIE PROBLEM ............................................... 29

FIGURE 3.3 ELECTRON CLOUD OSCILLATION IN SPHERICAL PARTICLES UNDER THE IMPACT

OF EXTERNAL EM WAVE .......................................................................................... 33

FIGURE 3.4 SCATTERING CROSS-SECTION OF SPHERICAL DIMER STRUCTURES WITH

DIFFERENT GAP DISTANCES ...................................................................................... 35

FIGURE 3.5 LSP EXCITATION AND SCATTERING ON THE SPHERICAL DIMERS .................... 36

FIGURE 3.6 TWO-STEP EM ENHANCEMENT MECHANISM OF RS ....................................... 38

FIGURE 4.1 SUMMARY FOR FABRICATION PROCESS OF SI BASE METAL COATED NP

STRUCTURES ............................................................................................................ 45

FIGURE 4.2 CLUSTERING OF NPS INDUCED BY CAPILLARY FORCES .................................. 47

FIGURE 4.3 LSP EXCITATION AND SIMULATED EF DISTRIBUTION ON AG COATED NP

DIMERS ..................................................................................................................... 49

FIGURE 5.1 PSEUDO FORCES IN CENTRIFUGAL MICROFLUIDICS PLATFORM. ...................... 59

FIGURE 5.2 PNEUMATIC PUMPING PRINCIPLE IN CENTRIFUGAL MICROFLUIDICS ............... 61

FIGURE 5.3 CAPILLARY ACTION IN CENTRIFUGAL MICROFLUIDICS ................................... 64

FIGURE 6.1 SAMPLE HANDLING AND FILTRATION USING THE CENTRIFUGAL PLATFORM ... 73

FIGURE 6.2 SERS STUDY OF “PURIFIED” AND “CLOGGED” REGIONS ............................... 75

FIGURE 6.3 QUANTITATIVE STUDY OF MELAMINE ............................................................ 76

FIGURE 6.4 FABRICATION OF MICROFLUIDIC DISCS AND EXPERIMENTAL SETUP ............... 80

FIGURE 6.5 MICROFLUIDIC PROCEDURE FOR NANOFILTRATION ....................................... 81

FIGURE 6.6 FLUID ARRANGEMENT DURING THE NANOFILTRATION PROCEDURE. .............. 84

FIGURE 6.7 SERS SPECTRA OF 0-20 PPM MELAMINE IN MILK ........................................... 85

FIGURE 6.8 AU NP SURFACE AFTER THE FILTRATION PROCESS. ....................................... 86

FIGURE 6.9 RAMAN SPECTRUM OF PURE MELAMINE IN POWDER FORM. ............................ 87

FIGURE 7.1 CENTRIFUGAL MICROFLUIDICS PLATFORM FOR SAMPLE HANDLING AND SERS

CHROMATOGRAPHY CONCEPT ON AU NP STRUCTURES ............................................ 95

FIGURE 7.2 MCR ANALYSIS OF SERS MAP OBTAINED ON AU COATED SHD NP

STRUCTURES AFTER THE CHROMATOGRAPHIC SEPARATION OF HUMAN URINE ......... 97

FIGURE 7.3 OPTIMIZATION OF NP DENSITY AND ITS INFLUENCE ON SERS

CHROMATOGRAPHY EFFICIENCY STUDIED USING HUMAN URINE ............................ 100

FIGURE 7.4 QUANTITATIVE ASSESSMENT OF PARACETAMOL IN HUMAN URINE USING SERS

CHROMATOGRAPHY TECHNIQUE ............................................................................. 102

FIGURE 7.5 FABRICATION PROCESS OF MICROFLUIDIC DISCS .......................................... 106

FIGURE 7.6 SCHEMATIC INTERPRETATION OF FLUID CONFIGURATIONS AT THE EQUILIBRIUM

CONDITIONS ........................................................................................................... 108

FIGURE 7.7 THE ROTATIONAL FREQUENCY PROFILE OF THE DISC FOR MICROFLUIDIC

PROCESSING OF THE SAMPLE .................................................................................. 110

FIGURE 7.8 CHARACTERIZATION OF NANOFILTRATION EFFECT THROUGH SURFACE

MORPHOLOGY STUDY OF AU COATED SHD NP STRUCTURES ................................. 112

FIGURE 7.9 TRANSFORMATION OF 2D SERS MAP INTO 1D SERS PROFILE .................... 114

FIGURE 7.10 THE PRINCIPLE OF MCR TECHNIQUE. ........................................................ 115

FIGURE 8.1 ILLUSTRATION OF THE FABRICATION PROCESS OF GOLD CAPPED GLASS

NANOPILLARS ......................................................................................................... 128

FIGURE 8.2 PHOTOGRAPH OF THE INTEGRATED SENSORS INTERFACED WITH A PRINTED

CIRCUIT BOARD. ..................................................................................................... 129

FIGURE 8.3 SERS CHARACTERIZATION AND OPTICAL PROPERTIES ................................. 130

FIGURE 8.4 EC CHARACTERIZATION .............................................................................. 132

FIGURE 8.5 DUAL SENSING OF PARACETAMO ................................................................. 133

FIGURE 8.6 FABRICATION OF NANOSTRUCTURED SURFACE ON FUSED SILICA.

ILLUSTRATIONS OF THE 4-INCH FUSED SILICA WAFERS WITH THE AL PATTERN ...... 134

FIGURE 8.7 CYCLIC VOLTAMMOGRAM OF 1 MM PAR IN PBS VS. AU PSEUDO-RE. ....... 135

FIGURE 8.8 CALIBRATION CURVE OF PARACETAMOL ..................................................... 136

LIST OF ABBREVIATIONS

RS – Raman Scattering

EM – Electromagnetic

SERS – Surface-enhanced Raman spectroscopy

EC – Electrochemistry

LSP – Localized surface plasmon

FDTD – Finite-difference time-domain

EF – Enhancement factor

MEF – Metal-enhanced fluorescence

NP – Nanopillar

RIE – Reactive ion etching

SEM – Scanning electron microscopy

PMMA – Poly (methyl methacrylate)

PSA – Pressure sensitive adhesive

CV – Cyclic voltammogram

WE – Working electrode

CE – Counter electrode

RE – Reference electrode

Ag – Silver

Au – Gold

Al – Alluminum

Cu – Copper

Si – Silicon

SiO2 – Fused Silica

SF6 – Sulphur hexafluoride

O2 – Oxygen

Chapter 1: Introduction

1

1 INTRODUCTION

The discovery of the surface-enhanced Raman scattering (RS) phenomenon can

be dated back to 1973, when an enormous Raman signal increase was observed for

pyridine molecules on electrochemically roughened Ag surface1. Today surface-

enhanced Raman spectroscopy (SERS) is a well-established and powerful technique for

studying molecular vibrations. The SERS method enables detection of tremendously

enhanced RS signals from molecules that are situated in the close vicinity (within a few

nm) of the plasmonic nanostructure surface2,3. Similar to the Raman spectroscopy

technique, the analysed molecular species are identified through their molecular

fingerprint spectra. It is well established that the boost in the RS signal is obtained via

(1) local concentration of the incident light at the nanoscale regions (“hot-spots”)

promoted by the localized surface plasmon (LSP) excitation4,5 (electromagnetic

enhancement), and (2) electronic band interaction between the analysed molecule and the

metal surface (chemical enhancement)6. Some specific cases for SERS based detection

of even single molecules have been already demonstrated7–9.

Only in recent decade, with the technological advancements in controlled micro-

and nanofabrication techniques, the SERS field started to gradually attract practical

interest in research environment as well as outside lab implementations10–16. The main

advantages of the SERS based systems are accounted for their molecular specificity, high

sensitivity and fluorescence quenching effect17,18. It was shown in multiple studies that

the SERS technique is suitable for examining the molecular reaction in electrochemistry

(EC) processes19,20, imaging and dynamic study of single molecule systems21,22. In


2

addition to that, a broad range of sensing applications on fluid and gaseous samples were

developed for detection of toxins for environmental23 and food safety research24,25,

pharmaceutical studies26, and medical diagnostics27,28. As a result, the technical

understanding behind the working mechanisms of the surface-enhanced RS effect and

essential requirements for plasmonic surfaces were interrogated and presented in

numerous theoretical and experimental reports. For developing reliable sensing

applications, SERS substrates require the following: (i) the fabrication procedure of a

SERS substrate should be scalable, reproducible and cost effective, (ii) display high

(>107) and uniformly distributed (>cm2) electromagnetic field enhancement factors (EF),

and lastly (iii) applicable for a broad range of LSP resonance modes15,16,29,30. On that

account, various highly ordered nanostructured SERS active surfaces with satisfactory

SERS performance were developed using mostly lithography based techniques31–33.

Despite all the progress in the SERS substrates fabrication field, current SERS sensing

methodologies are not universal, lack efficiency and reproducibility for analysing

complex, multicomponent real-life samples compared to e.g. existing techniques such as

high-performance liquid chromatography (HPLC)34,35, capillary electrophoresis36,37

integrated detectors, lateral flow based assays38,39 or traditional Raman spectroscopy40,41.

The main technical challenge for SERS based detection of analyte molecules in

complex media such as blood plasma, urine, saliva, milk, juice, waste water, etc. is related

to their chemically rich environment23,27,42. Since the phenomenon is highly surface

sensitive, an ideal SERS measurement is necessitating relatively pure sample

environment on the active region. In general terms, macromolecular structures such as

proteins, cells, fats tend to block electromagnetic “hot-spots” which significantly limits

further SERS-based detection of analyte compounds43. Moreover, on some occasions the

sample fluid may contain multiple molecular species with varying affinities towards

nanostructured metal surface. Due to competitive binding of these different components


3

on a SERS active region44, the limit of detection for distinct analyte species can be greatly

reduced.

A number of sample pre-treatment and handling methods to overcome above

mentioned practical issues for SERS based applications on complex suspensions have

already been demonstrated. For example, functionalization of the SERS active surface43,

analytical liquid-liquid extraction technique45,46, filtration on porous membranes47,48,

analyte enrichment procedures49, integration of microfluidic devices50–52 and

incorporation of external chromatographic platforms such as HPLC and thin-layer

chromatography (TLC)53–55 are the most prominent sample handling techniques for

SERS. Nevertheless, the current integrated SERS based platforms are application

specific, time consuming, lack sensitivity or do not provide quantitative information.

1.1 Aim of the PhD Project

This PhD study is a part of the HERMES-High Exponential Rise in Miniaturized

cantilever-like Sensing project funded by European Research Council (ERC). The main

objective of the project is to solve fundamental challenges of SERS based sensing

applications in real-life complex samples like blood, milk and saliva. To accomplish this,

the SERS measurements ideally should be (i) reliable, (ii) sensitive, (iii) reproducible and

(iv) high-throughput. These sensing characteristics can be achieved by utilizing high

performance, nanopillar-based SERS substrates and controlled sample pre-

treatment\deposition techniques. The state-of-the-art metal capped nanopillar (NP) SERS

substrates developed by senior researcher Michael Stenbæk Schmidt in the Nanoprobes

group are potentially suitable for practical applications. Thus, in this PhD project, the

contribution of controlled sample pre-treatment and dosage techniques for SERS based

sensing are examined, and the proof-of-concept studies are performed on the Au capped


4

NP arrays. Accordingly, these fluid manipulation steps are accomplished using a

centrifugal microfluidics platform.

Specific objectives of the PhD project are the following: (i) design, optimize and

fabricate polymeric centrifugal microfluidic discs for desired fluid handling steps, (ii)

integrate the NP-based SERS active surfaces into the centrifugal microfluidics platform

and (iii) demonstrate potential applicability of the integrated platform using well-known

sensing examples in complex sample solutions. Lastly, it should be emphasized that the

proposed solutions in the SERS based sensing system are aimed to be versatile,

automated, free of human interaction, and more importantly without any surface

functionalization.

1.2 Structure of the Thesis

This introductory part of the thesis (Chapter 1) is followed by 4 chapters that

provide theoretical background for this PhD project. Namely, in Chapter 2 (Raman

Effect), the fundamental concept behind the SERS phenomenon, inelastic light scattering

effect (Raman scattering) for probing the vibrational modes of molecules is given. Further,

the surface-enhancement of Raman scattering signal is elaborated through the

electromagnetic enhancement mechanism in Chapter 3 (Surface-Enhanced Raman

Scattering). The Au coated NP SERS substrates which have been utilized throughout the

project are described in Chapter 4 (Plasmonic NP Structures). Lastly, the fluid handling

platform, i.e. centrifugal microfluidics, is presented in Chapter 5 (Centrifugal

Microfluidics). The results obtained during the PhD project are discussed in Chapters 6-

8. In Chapter 6 (Nanopillar filters for SERS), a facile and simple method to eliminate

macromolecular clusters through the wicking effect of fluids and detection of small

analyte molecules in complex sample medium is described. The multicomponent analyte

detection in chemically rich sample environment is introduced in Chapter 7 (NP


5

Structures for SERS Chromatography) using a similar technique as in Chapter 6. An

alternative SERS based sensing approach is demonstrated in Chapter 8 (Dual-Functional

EC and SERS Sensing). Here the two techniques, i.e. electrochemical detection and SERS

based qualitative analysis, were implemented for sample analysis. Finally, conclusions

and an overview of main achievements of the PhD project are summarized in Chapter 9

(Concluding Remarks).

References

(1) Fleischmann, M.; Hendra, P. J.; McQuillan, A. J. Chem. Phys. Lett. 1974, 26 (2),

163–166.

(2) Otto, A.; Mrozek, I.; Grabhorn, H.; Akemann, W. J. Phys. Condens. Matter

1992, 4 (5), 1143–1212.

(3) Kneipp, K.; Kneipp, H.; Itzkan, I.; Dasari, R. R.; Feld, M. S. J. Phys. Condens.

Matter 2002, 14 (18), 202.

(4) Campion, A.; Kambhampati, P. Chem. Soc. Rev. 1998, 27 (4), 241.

(5) Stiles, P. L.; Dieringer, J. A.; Shah, N. C.; Van Duyne, R. P. Annu. Rev. Anal.

Chem. 2008, 1 (1), 601–626.

(6) Le Ru, E. C.; Blackie, E.; Meyer, M.; Etchegoin, P. G. J. Phys. Chem. C 2007,

111 (37), 13794–13803.

(7) Camden, J. P.; Dieringer, J. A.; Wang, Y.; Masiello, D. J.; Marks, L. D.; Schatz,

G. C.; Van Duyne, R. P. J. Am. Chem. Soc. 2008, 130 (38), 12616–12617.

(8) Nie, S. Science (80-. ). 1997, 275 (5303), 1102–1106.

(9) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.;

Feld, M. S. Phys. Rev. Lett. 1997, 78 (9), 1667–1670.

(10) Haes, A. J.; Haynes, C. L.; McFarland, A. D.; Schatz, G. C.; Van Duyne, R. P.;

Zou, S. MRS Bull. 2005, 30 (5), 368–375.

(11) Cialla, D.; März, A.; Böhme, R.; Theil, F.; Weber, K.; Schmitt, M.; Popp, J.

Anal. Bioanal. Chem. 2012, 403 (1), 27–54.

(12) Petry, R.; Schmitt, M.; Popp, J. ChemPhysChem 2003, 4 (1), 14–30.


6

(13) Haynes, C. L.; Yonzon, C. R.; Zhang, X.; Van Duyne, R. P. J. Raman Spectrosc.

2005, 36 (6–7), 471–484.

(14) Halvorson, R. A.; Vikesland, P. J. Environ. Sci. Technol. 2010, 44 (20), 7749–

7755.

(15) Moskovits, M. In Surface-Enhanced Raman Scattering; Springer Berlin

Heidelberg; pp 1–17.

(16) Sharma, B.; Frontiera, R. R.; Henry, A.-I.; Ringe, E.; Van Duyne, R. P. Mater.

Today 2012, 15 (1–2), 16–25.

(17) Le Ru, E. C.; Etchegoin, P. G. In Principles of Surface-Enhanced Raman

Spectroscopy; Elsevier, 2009; pp 1–27.

(18) Hildebrandt, P.; Stockburger, M. J. Phys. Chem. 1984, 88 (24), 5935–5944.

(19) Tian, Z.-Q.; Ren, B. Annu. Rev. Phys. Chem. 2004, 55 (1), 197–229.

(20) Wu, D.-Y.; Li, J.-F.; Ren, B.; Tian, Z.-Q. Chem. Soc. Rev. 2008, 37 (5), 1025.

(21) Kleinman, S. L.; Ringe, E.; Valley, N.; Wustholz, K. L.; Phillips, E.; Scheidt, K.

A.; Schatz, G. C.; Van Duyne, R. P. J. Am. Chem. Soc. 2011, 133 (11), 4115–4122.

(22) Zhang, R.; Zhang, Y.; Dong, Z. C.; Jiang, S.; Zhang, C.; Chen, L. G.; Zhang, L.;

Liao, Y.; Aizpurua, J.; Luo, Y.; Yang, J. L.; Hou, J. G. Nature 2013, 498 (7452), 82–86.

(23) Schlücker, S. Angew. Chemie Int. Ed. 2014, 53 (19), 4756–4795.

(24) Craig, A. P.; Franca, A. S.; Irudayaraj, J. Annu. Rev. Food Sci. Technol. 2013, 4

(1), 369–380.

(25) Zheng, J.; He, L. Compr. Rev. Food Sci. Food Saf. 2014, 13 (3), 317–328.

(26) Cînta Pînzaru, S.; Pavel, I.; Leopold, N.; Kiefer, W. J. Raman Spectrosc. 2004,

35 (5), 338–346.

(27) Cotton, T. M.; Kim, J.-H.; Chumanov, G. D. J. Raman Spectrosc. 1991, 22 (12),

729–742.

(28) Qian, X.; Peng, X.-H.; Ansari, D. O.; Yin-Goen, Q.; Chen, G. Z.; Shin, D. M.;

Yang, L.; Young, A. N.; Wang, M. D.; Nie, S. Nat. Biotechnol. 2008, 26 (1), 83–90.

(29) Sharma, B.; Fernanda Cardinal, M.; Kleinman, S. L.; Greeneltch, N. G.;

Frontiera, R. R.; Blaber, M. G.; Schatz, G. C.; Van Duyne, R. P. MRS Bull. 2013, 38 (8),

615–624.

(30) Moskovits, M. Rev. Mod. Phys. 1985, 57 (3), 783–826.


7

(31) Fan, M.; Andrade, G. F. S.; Brolo, A. G. Anal. Chim. Acta 2011, 693 (1–2), 7–

25.

(32) Wang, A.; Kong, X. Materials (Basel). 2015, 8 (6), 3024–3052.

(33) Procházka, M. Surface-Enhanced Raman Spectroscopy; Biological and Medical

Physics, Biomedical Engineering; Springer International Publishing: Cham, 2016.

(34) Swadesh, J. K. HPLC: practical and industrial applications; CRC Press, 2000.

(35) Anderson, D. J. Anal. Chem. 1999, 71 (12), 314–327.

(36) Shintani, H. Handbook of capillary electrophoresis applications; Springer

Science & Business Media, 2012.

(37) Grossman, P. D.; Colburn, J. C. Capillary electrophoresis: Theory and practice;

Academic Press, 2012.

(38) Fenton, E. M.; Mascarenas, M. R.; Lopez, G. P.; Sibbett, S. S. ACS Appl. Mater.

Interfaces 2009, 1 (1), 124–129.

(39) Posthuma-Trumpie, G. A.; Korf, J.; van Amerongen, A. Anal. Bioanal. Chem.

2009, 393 (2), 569–582.

(40) Schrader, B. Infrared and Raman spectroscopy: methods and applications; John

Wiley & Sons, 2008.

(41) Lewis, I. R.; Edwards, H. Handbook of Raman spectroscopy: from the research

laboratory to the process line; CRC Press, 2001.

(42) Bantz, K. C.; Meyer, A. F.; Wittenberg, N. J.; Im, H.; Kurtuluş, Ö.; Lee, S. H.;

Lindquist, N. C.; Oh, S.-H.; Haynes, C. L. Phys. Chem. Chem. Phys. 2011, 13 (24), 11551.

(43) Sun, F.; Hung, H.; Sinclair, A.; Zhang, P.; Bai, T.; Galvan, D. D.; Jain, P.; Li,

B.; Jiang, S.; Yu, Q. Nat. Commun. 2016, 7, 13437.

(44) Yu, W. W.; White, I. M. Analyst 2013, 138 (13), 3679.

(45) Morelli, L.; Zór, K.; Jendresen, C. B.; Rindzevicius, T.; Schmidt, M. S.; Nielsen,

A. T.; Boisen, A. Anal. Chem. 2017, 89 (7), 3981–3987.

(46) Cheng, J.; Wang, S.; Su, X. O. PLoS One 2014, 9 (9).

(47) Yu, W. W.; White, I. M. Manuscript, Author 2012, 137 (5), 1168–1173.

(48) Kim, A.; Barcelo, S. J.; Williams, R. S.; Li, Z. Anal. Chem. 2012, 84 (21), 9303–

9309.


8

(49) Morelli, L.; Andreasen, S. Z.; Jendresen, C. B.; Nielsen, A. T.; Emnéus, J.; Zór,

K.; Boisen, A. Analyst 2017, 142 (23), 4553–4559.

(50) Lee, M.; Lee, K.; Kim, K. H.; Oh, K. W.; Choo, J. Lab Chip 2012, 12 (19), 3720.

(51) Strehle, K. R.; Cialla, D.; Rösch, P.; Henkel, T.; Köhler, M.; Popp, J. Anal.

Chem. 2007, 79 (4), 1542–1547.

(52) Ackermann, K. R.; Henkel, T.; Popp, J. ChemPhysChem 2007, 8 (18), 2665–

2670.

(53) Subaihi, A.; Trivedi, D. K.; Hollywood, K. A.; Bluett, J.; Xu, Y.; Muhamadali,

H.; Ellis, D. I.; Goodacre, R. Anal. Chem. 2017, 89 (12), 6702–6709.

(54) Yu, W. W.; White, I. M. Analyst 2013, 138 (13), 3679.

(55) Lin, M.; He, L.; Awika, J.; Yang, L.; Ledoux, D. R.; Li, H.; Mustapha, A. J.

Food Sci. 2008, 73 (8), T129–T134.

Chapter 2: Raman Effect

9

2 RAMAN EFFECT

Common and broadly applied optical spectroscopy techniques (e.g. Raman,

Infra-red, absorption spectroscopy) are focused on analysis of light-matter interactions.

The phenomenon of light scattering is one of the typical outcomes of such interactions.

The information about vibrational modes of molecules, crystal orientation and other

morphological parameters can be retrieved from the scattered light1,2. A straightforward

reason for an electromagnetic (EM) wave to experience scattering, which implies change

of its propagation direction, is related to optical inhomogeneity of the medium. Thus, time

dependent fluctuations of local refractive index results in scattering of incident light3. In

a general form, refractive index 𝒏 of a medium is given as:

where 𝑴 is concentration of molecules and 𝜶 is the local polarizability of the medium.

According to Equation (2.1), a time-dependent change in refractive index can be

originating from a varying local molecule concentration or polarizability. Particularly the

last reason, fluctuations in polarizability, causes the incident EM wave to experience

Raman scattering (RS) which contains information about vibrational modes of molecules

in the medium.

In this chapter, one of the core element of this multidisciplinary PhD project, RS

of light will be presented. In this way, physical background of the phenomena, current

status, main advantages and disadvantages will be discussed. Before presenting a

quantitative definition of RS using a classical approach, intermolecular oscillations and

𝑛2 = 1 + 𝑀𝛼, (2.1)


10

their relationship with the polarizability will be described. To keep discussions brief, a

medium containing only one sort of molecules will be considered.

2.1 Molecular Vibration Modes

In classical theory, the simplest way of defining intermolecular vibrations is

“string” approximation where atoms in a molecule are assumed to interact according to

Hooke’s law. In addition to that, non-harmonic behavior of “string” constant 𝒌 in

agreement with semi-empirical models (Morse model, Lennard-Jones model) can be

implemented to depict a more adequate picture of molecular systems4. The classical

harmonic oscillator and more precise semi-empirical models cannot fully describe the

interatomic interactions. These are rather complex in nature since they depend on the

type of the chemical bond (ionic, metallic or covalent) and various quantum effects like

Figure 2.1 Schematic illustration of a simplified diatomic molecule. The atoms are illustrated

as spheres attached through a “string” with a string constant 𝒌. The spatial positions of atoms with

respect to reference system O', relative distance between atoms and their masses are indicated

through vectors 𝒓ሬԦ𝟏, 𝒓ሬԦ𝟐, 𝜟𝒓ሬԦ, 𝒎𝟏 and 𝒎𝟐 respectively.


11

repulsion force due to Pauli exclusion principle, quantization of energy levels which

cannot be represented via classical approaches5,6.

A straightforward example of a classical model of a diatomic molecule is shown

in Figure 2.1. For clarity of further derivations, only harmonic oscillations (linear force

response of the “string”) will be taken into account. In fact, this estimation is valid when

assuming small fluctuations in non-harmonic molecular oscillations. Further, the

equilibrium condition of the system will be defined at constant relative distance of 𝜟𝒓ሬԦ𝟎

between the atoms. Hence, by applying Newtons law, a differential equation of motion

can be obtained:

By combining Equation (2.2)-(2.4) the final form of the differential motion equation can

be reduced to:

A general solution for the second order differential Equation (2.5), which is commonly

referred in literature as equation for vibrational motion7, can in scalar form be expressed

as:

𝑚1

𝑑2𝑟Ԧ1𝑑𝑡2

= 𝑘(𝛥𝑟Ԧ − 𝛥𝑟Ԧ0), (2.2)

𝒎𝟐

𝒅𝟐𝒓ሬԦ𝟐

𝒅𝒕𝟐= −𝒌(𝜟𝒓ሬԦ − 𝜟𝒓ሬԦ𝟎) , (2.3)

𝜟𝒓ሬԦ = 𝒓ሬԦ𝟐 − 𝒓ሬԦ𝟏 . (2.4)

𝑑2(𝛥𝑟Ԧ − 𝛥𝑟Ԧ0)

𝑑𝑡2= −

𝑘(𝑚1 + 𝑚2)

𝑚1𝑚2

(𝛥𝑟Ԧ − 𝛥𝑟Ԧ0) . (2.5)

𝑄 = ‖𝛥𝑟Ԧ − 𝛥𝑟Ԧ0‖ = 𝐴𝑒𝑖Ω𝑡 + 𝐵𝑒−𝑖Ω𝑡 , (2.6)


12

where 𝑨 and 𝑩 are the amplitudes and Ω is the angular frequency of vibrational motion

and it is defined as:

Now, assuming a molecule which contains 𝑵 atoms, a common way to define

oscillation between atoms is to utilize superposition principle. This suggests that each

interatomic vibration in a molecule can be expressed in terms of independent oscillations

or “normal modes” with a natural vibrational frequency Ω𝒊 (see Equation (2.7)). The total

amount of modes (3𝑵 − 6) which represents degree of freedom of the molecule system

is directly linked to the number of atoms in that molecule. For instance, a water molecule

has 3 atoms, hence it has 3 degrees of freedom (or “normal modes”). The independent

vibrational modes of a water molecule which are bending, symmetrical and asymmetrical

stretching are shown in Figure 2.2. Thus, for a molecule consisting of 𝑵 atoms, by

Ω2 =𝑘(𝑚1 + 𝑚2)

𝑚1𝑚2 . (2.7)

Figure 2.2 Various “normal” vibrational modes of the water molecule. The arbitrary

oscillatory motion of a water molecule can be given as superposition of these three independent

modes (a) symmetrical, (b) asymmetrical stretching and lastly (c) bending.


13

applying superposition principle, the vibration between 2 atoms can be analytically

expressed in the following way:

The final obtained expression for total vibration between 2 atoms, Equation (2.8),

describes fully the oscillating molecule system. As a remark, the Equation (2.8) was

derived by assuming harmonic interactions between atoms. Nevertheless, for upcoming

subsection, this assumption is sufficient to depict the relation between polarizability of a

molecular system and vibrational “normal” modes.

2.2 Polarizability of Molecules

An essential quantitative parameter describing the interaction between matter

and EM waves (or light) is the polarizability of the medium. Polarizability defines the

local charge distribution or more precisely the dipole moment in a unit volume

(polarization) under the influence of electric field. In general terms, the molecules which

form the matter can be divided into two species, (i) polar molecules which have natural

dipole moment without any external impact and (ii) non-polar molecules which have

negligible dipole moment8. An illustrative example is given in Figure 2.3a, b where polar

H2O and non-polar CO2 molecules are shown. Thus, because of its asymmetrical charge

distribution, water molecule possesses a dipole moment. Contrary, the CO2 molecule with

a linear arrangement has negligible amount of total electrical dipole moment9. However,

taking into account the chaotic motion of polar molecules in the substances, the resulting

effect of electric dipole in terms of polarization is insignificant as well.

𝑄𝑡 = ∑ 𝑄𝑗 =

3𝑁−6

𝑗=1

∑ (𝐴𝑗

3𝑁−6

𝑗=1

𝑒𝑖Ω𝑗𝑡 + 𝐵𝑗𝑒−𝑖Ω𝑗𝑡) . (2.8)


14

The behavior of polar and non-polar molecules under an externally applied

electric field is different. For simplicity, assuming only uniformly applied fields, the

molecules with naturally present dipole moment would align with respect to the field

direction. In other words, the main contribution to the polarization of the substance is

emerging from the dipole moments (polar molecules) which adjust their orientation with

the help of Coulomb forces. On the other hand, the non-polar molecules can provide

polarization of the matter as well. As it is presented in Figure 2.3c, for a neutrally charged

atom, the formation of dipole moments under an external field relies on spatial shifts of

localized charges. In this way, by pulling apart opposite charges, electron cloud and

positively charged nuclei, the dipole moment and consequently the polarization of the

Figure 2.3 Polarizability in molecular scale. (a) Natural dipole moment of water molecule due

to its linear asymmetry and charge distribution. (b) Non-polar CO2 molecule as a result of its

linear symmetry has negligible dipole moment. (c) Principle of formation of dipole moment on

non-polar molecules or atoms under the effect of an electric field.


15

substance is achieved8. A similar effect will be discussed in the next chapter where the

impact of EM wave on plasmonic structures is introduced. Nevertheless, regardless of the

polarity of the molecule, the dipole moment has a linear correspondence with the

externally applied electric field:

where 𝒑ሬሬԦ is the dipole moment, 𝜶 is the polarizability and 𝑬ሬሬԦ is the external electric field.

One should note that the equality is valid only if the external electric field is not strong

enough to result in structural changes of the molecule.

The Equation (2.9) is not precise enough in the sense that a time dependent

fluctuation of the polarizability function is not defined. For most of the applications, it

might be sufficient to consider the time averaged response from the electric field. But as

it was stated earlier in this chapter, the RS is caused by local time dependent polarizability

fluctuations of the molecule. The core understanding behind such time dependent changes

in polarizability relies on the dynamic nature of molecules. More accurately, the

intermolecular oscillations, which implies constantly changing distances between atoms

and as a result local charge distributions suggest distinct response (polarizability) to the

external field. Since the Equation (2.8) is a known time dependent function, the unknown

function 𝜶(𝒕) can be decomposed around equilibrium position (‖𝜟𝒓ሬԦ − 𝜟𝒓ሬԦ𝟎‖ = 0) using

Taylor series:

Now, assuming that the proposed Taylor series is converging, the accuracy of the model

will be defined by the number of elements which would be considered. For our purpose,

𝑝Ԧ = 𝛼𝐸ሬԦ , (2.9)

𝛼(𝑡) = ∑𝜕𝑙𝛼

𝜕𝑄𝑡𝑙

∞

𝑙=1

(0)(𝑄𝑡)𝑙

𝑙! . (2.10)


16

first two terms are sufficient to illustrate the principle of RS. Thus, for the following

expressions, the time dependent polarizability function:

will be used.

2.3 Rayleigh and Raman Scattering

In previous subsections, intermolecular vibrations and their relationship with the

polarizability of a molecule was presented. Equation (2.9) and (2.11) provide an essential

and sufficient base for classical EM interpretation of RS. The phenomenon involves an

external illumination of the molecules and subsequent scattering of the light. For clarity,

as an incident illumination source, monochromatic and coherent EM wave will be taken

into account. The overall light scattering process is summarized in Figure 2.1Figure 2.4.

As a result, the following relation can be retrieved:

Hence, the final form can be reduced to:

where 𝑬𝟎 is the electric field amplitude and 𝝎𝟎 is the angular frequency of the incident

EM wave. It’s a well-known fact that an oscillating electric dipole 𝒑ሬሬԦ(𝒕) acts as a source

of light. Moreover, the oscillation frequency of the dipole is directly linked to the

𝛼(𝑡) = 𝛼(0) +𝜕𝛼

𝜕𝑄𝑡

(0) ∑ (𝐴𝑗

3𝑁−6

𝑗=1

𝑒𝑖Ω𝑗𝑡 + 𝐵𝑗𝑒−𝑖Ω𝑗𝑡) , (2.11)

‖𝑝Ԧ‖ = ‖𝛼𝐸ሬԦ‖ = (𝛼(0) +𝜕𝛼

𝜕𝑄𝑡

(0) ∑ (𝐴𝑗

3𝑁−6

𝑗=1

𝑒𝑖Ω𝑗𝑡 + 𝐵𝑗𝑒−𝑖Ω𝑗𝑡)) (𝐸0𝑒𝑖𝜔0𝑡). (2.12)

‖𝑝Ԧ(𝑡)‖ = 𝐸0𝛼(0)𝑒𝑖𝜔0𝑡 + 𝐸0

𝜕𝛼

𝜕𝑄𝑡

(0) ∑ (𝐴𝑗

3𝑁−6

𝑗=1

𝑒𝑖𝑡(𝜔0+Ω𝑗) + 𝐵𝑗𝑒𝑖𝑡(𝜔0−Ω𝑗)), (2.13)


17

frequency of the emitted light and in a similar manner, the amplitude of dipole oscillation

defines the power of the total emitted light. In such manner, the first term of Equation

(2.13) represents the Rayleigh scattering of the incident light where the scattered light has

exactly the same angular frequency as compared to the incident EM wave. Further, the

last term of the equation represents the inelastic RS of the light with divergent angular

frequency of the dipole oscillation. Commonly in literature, the oscillation with higher

frequency 𝝎𝟎 + Ω𝒋 is referred to as anti-Stokes scattering while the lower frequency

𝝎𝟎 − Ω𝒋 is Stokes scattering of incident light. It is evident that both Stokes and anti-

Stokes scattering of the light contains spectroscopically rich information about the

vibrational modes of the molecule. However, the RS of light relies on certain conditions.

First of all, the intensity of Rayleigh scattering dominates over RS and it’s a typical event

for all sort of molecules. This can be correlated with the fact that all types of molecules

Figure 2.4 Schematic illustration of the scattering process. As a result of illumination of a

Raman active molecule (melamine) with incident EM wave with angular frequency of 𝝎𝟎, three

distinct types of scattering processes take place. Those are Rayleigh, Stokes and anti-Stokes

scattering of the light.


18

can be polarized to some extent. On the other hand, less probable event RS (only 10-5 part

of incident light) requires non-zero polarizability gradient (around the equilibrium

position) with respect to intermolecular vibrations10:

If the specification (selection rule) in Equation (2.14) is fulfilled, the molecule is called

Raman active. As an example, a spectrum of the Raman active molecule melamine

recorded at 785 nm laser illumination is shown in Figure 2.5. Here, Stokes scattering on

melamine crystal structures (intensity) and accordingly the intermolecular vibration

modes are presented through Raman shift which gives relative deviation from the laser

wavelength (𝞴𝒍𝒂𝒔𝒆𝒓) in terms of wavenumbers:

𝜕𝛼

𝜕𝑄𝑡

(0) ≠ 0 . (2.14)

Figure 2.5 Raman spectrum of melamine recorded at 785 nm laser illumination.


19

where 𝞴𝟏 is the wavelength at which scattering intensity is measured.

As it was mentioned previously in this chapter, the classical approach is not

adequate to give a full picture of the RS phenomenon. The classical approach turns out to

be sufficient to predict Rayleigh scattering accurately. However, the classical model

(Equation (2.13)) is adequate only for qualitative definition of RS, which means that it

can be utilized to determine the frequency of the scattered EM wave. Further, the classical

assumption of RS intensity, in other words quantitative evaluation of the process is valid

only for limited cases. Moreover, Equation (2.13) describes only the vibrational RS

neglecting the rotational modes. The pure origin of those problems lies on various

mechanisms such as restriction rules for quantized energy level transitions, population of

energy states and discretization of rotational modes which are not present in classical

approach10.

2.4 Raman Spectroscopy

Nowadays, Raman spectroscopy is a widely employed technique. Commonly,

the spectroscopic tool is utilized for sample analysis in various applications of

biomedical, environmental, pharmaceutical and semi-conductor research fields. The

technique enables one to identify and at the same time quantify certain analytes using the

fingerprint spectra in a broad range specimens (liquids, solids and gasses) which makes

the Raman spectroscopy a highly specific and versatile tool. Moreover, the method does

not require expertise in sample preparation and it is also possible to acquire spectral

response within seconds from very small sample amounts (for liquids ~µl). According to

recent studies, ~µM detection limit of analyte molecules with high Raman cross-section

𝑅𝑎𝑚𝑎𝑛 𝑠ℎ𝑖𝑓𝑡 = (1

𝜆𝑙𝑎𝑠𝑒𝑟−

1

𝜆1) , (2.15)


20

is possible to achieve9,10. Nevertheless, for some application cases, the technique is

incapable of performing a sample analysis. The main disadvantages of Raman

spectroscopy can be addressed to the nature of light matter interactions. First of all, along

with the scattering process, the incident light can also excite electronic states of molecules

in specimen which eventually would result in fluorescence emission. Taking into account

that the RS effect is weak (only 10-5 part of incident light), the broad fluorescence

background signal would cover the spectroscopic fingerprints of the molecule. For

example, a broad fluorescence background signal is observed in Raman measurements of

alkaloids on plant cell samples using an excitation wavelength of 365 nm. Even though

the analyte molecule is not fluorescent under 365 nm excitation wavelength, because of

the chemically rich environment of the biological sample, fluorescent background signal

from macromolecules are prohibiting the analysis of the analyte molecule12. One way to

overcome this issue is to utilize different excitation laser sources for example in near

infra-red region. However, the Raman scattering intensity (𝑰𝑹𝒂𝒎𝒂𝒏) is highly dependent

on excitation wavelength (𝞴𝑳𝒂𝒔𝒆𝒓)13:

where 𝑰𝑳𝒂𝒔𝒆𝒓 is the intensity of incident laser. Optionally this can be compensated by

increasing the excitation laser intensity. Unfortunately, in most cases, this leads to

destructive effects on the molecular structure of analytes through local heating of the

laser.

References

(1) Chylek, P.; Ramaswamy, V.; Ashkin, A.; Dziedzic, J. M. Appl. Opt. 1983, 22

(15), 2302.

𝐼𝑅𝑎𝑚𝑎𝑛 =𝐼𝐿𝑎𝑠𝑒𝑟

𝜆𝐿𝑎𝑠𝑒𝑟4 , (2.16)


21

(2) Pecora, R. Dynamic light scattering: applications of photon correlation

spectroscopy; Springer Science & Business Media, 2013.

(3) Stockmayer, W. H. J. Chem. Phys. 1950, 18 (1), 58–61.

(4) John R. Ferraro, Kazuo Nakamoto, C. W. B. Introductory Raman spectroscopy;

Boston : Academic Press, 2003.

(5) Brown, J. M.; Carrington, A. Rotational spectroscopy of diatomic molecules;

Cambridge University Press, 2003.

(6) Hollas, J. M. Modern spectroscopy; John Wiley & Sons, 2004.

(7) Thomson, W. Theory of vibration with applications; CRC Press, 1996.

(8) Griffiths, D. J. Introduction to Electrodynamics; Prentice Hall, 1999.

(9) Nir, S.; Adams, S.; Rein, R. J. Chem. Phys. 1973, 59 (6), 3341–3355.

(10) Long, D. A. The Raman Effect; John Wiley & Sons, Ltd: Chichester, UK, 2002.

(11) Schrader, B. Infrared and Raman spectroscopy: methods and applications; John

Wiley & Sons, 2008.


(13) Cox, A. J.; DeWeerd, A. J.; Linden, J. Am. J. Phys. 2002, 70 (6), 620–625.


22

Chapter 3: Surface-Enhanced Raman Scattering

23

3 SURFACE-ENHANCED RAMAN

SCATTERING

In Chapter 1 classical depiction of RS phenomenon and a brief overview of the

spectroscopic tool which is based on this effect was given. As it was mentioned, the

capability of Raman spectroscopy to record fingerprint spectrum of molecules in the

specimen and in this way distinguishing them among each other has attracted a great

interest for various practical implementations. For instance, Raman spectroscopy is a

suitable tool for electrochemistry (EC) applications to monitor the surface reactions on

conductive electrodes. However, in some cases, the detection limit which the Raman

technique can provide is not sufficient for detecting the molecular absorption events on

electrode surfaces. One way to overcome this is to utilize porous electrode structures with

higher surface area. Thus, during the EC treatment of the sample, the surface

concentration of distinct absorbed molecules on the electrode (𝑴) would be increased1.

It is obvious that the Raman signal for a specific molecule type is linearly dependent on

number of excited molecules or more precisely the amount of Raman scatterers in the

laser spot. Using the Equation (2.13), in a simple manner, the Raman scattering intensity

can be given in the following way:

where ‖𝑬ሬሬԦ𝟎‖𝟐

is related to intensity of the laser illumination and (𝝏𝜶

𝝏𝑸𝒕) is associated with

the Raman activity (cross-section) of the molecule. A similar study was carried out by

𝐼𝑅𝑎𝑚𝑎𝑛~𝑀 ‖𝐸ሬԦ0‖2

(𝜕𝛼

𝜕𝑄𝑡)

2

, (3.1)


24

Fleischmann et al., where roughened Ag electrodes were employed for EC measurements

of pyridine molecules2. According to their observations, enormous boost of Raman signal

from pyridine molecules on electrode surface was not correlated with the increased

surface area. This was the first experimental report on the surface-enhanced RS effect on

roughened Ag surface which suggested a unique boosting mechanism. Over the past

decades, multiple theoretical and experimental investigations on understanding the

working principle of this surface phenomenon on different roughened metal surfaces were

conducted3–5. As a result, the boost of RS is explained in terms of two contributing

fundamental principles. First, the EM enhancement mechanism which results in a local

increase of the electric field, ‖𝑬ሬሬԦ𝟎‖𝟐

term in Equation (3.1). Further, a chemical

enhancement mechanism which occurs due to electron band interaction between the

analyzed molecule and the metal surface (e.g. Ag), (𝝏𝜶

𝝏𝑸𝒕) term in Equation (3.1).

Particularly the last mechanism indicates that the analyzed molecule and the metal surface

cannot be treated as two isolated systems. Instead, the bound system should be

characterized as one individual system with unique polarizability and Raman cross-

section. In fact, this mechanism is significantly dependent on the molecule and the

interacting metal type6. However, for the broad variety of roughened conducting surfaces,

the main contribution for enhancement is achieved through the EM mechanism7. For this

reason, in upcoming sections, the basic concepts of surface-enhanced RS phenomenon in

terms of EM boosting principle will be presented.

3.1 EM Waves in Conducting Environment

It should be noted that in consonance with the experimental observations, the

surface-enhanced RS phenomenon was observed on roughened surfaces (in most of the

cases nanostructured) of conducting noble metals such as Ag, Au, Al and Cu7. Therefore,


25

before giving a theoretical explanation on EM enhancement mechanism, the behavior of

the conductive materials under the impact of external EM wave should be described. As

it was pointed out in Section 2.2, the response of matter under an electric field is

characterized in terms of molecular polarizability when dealing with microscale systems.

However, the dielectric permittivity function, which defines the polarization (dipole

moment in unit volume) is more suitable for interpretation of macroscale systems like

roughened Ag surface. It should be emphasized that the dielectric function was initially

introduced to define insulators that have limited capacity for charge transfer. Moreover,

under a static electric field, the polarization would be linearly dependent on the amplitude

Figure 3.1 Drude model for electron transfer in conducting materials. The positively charged

metal ions are illustrated as blue spheres and the charge carriers (electrons) of the conductor is

presented as red spheres. The charge transfer which proposes net collective motion of electrons

is facilitated by an externally applied electric field 𝑬. Further, the dissipative interaction between

electrons and ions is characterized through a characteristic collision frequency parameter.


26

of the field. Indeed, under the effect of static electric field, the dielectric permittivity is

not applicable for conductor materials due to presence of the unbound electrons.

Nevertheless, the dielectric permittivity function can be associated with if the applied

field is oscillating (EM wave). Under this circumstance, the spatial migration of charge

carriers in conductors would be restricted due to time dependent changes in the electric

field direction8.

Generally, a conductive material can be classified as a system of atoms which

consists of unbound or more specifically “loose electrons” which are easily distorted

under any external EM influence. In this way, the conductivity of e.g. metals is explained

as a flow of electrons which are easily brought into motion by means of an applied electric

field8. In the early 1900’s the theoretical investigation of charge transport in conductive

materials attracted great interest. The simplest and also accurate classical representation

of electron motion in metals under an external perturbation can be achieved through the

Drude model (Figure 3.1). Thus, according to the model, the conductor is assumed to be

arrangement of positively charged metal ions, which are firmly positioned in a periodic

configuration and a uniformly distributed mobile electron cloud. Additionally, the chaotic

nature of electron motion is described via a characteristic parameter of a material, the

relaxation time. In summary, the relaxation time defines statistically averaged collision

frequency (𝞿) between an electron in motion and immobile positively charged ions in the

metal. It is obvious that the relaxation time would be defined by multiple factors like

density, charge and mass of the ions and electrons in the metal8–10. Taking into account

all the aspects, the equation of motion for an electron under EM wave, according to Drude

model can be expressed as:

𝑚𝑒

𝑑2𝑟Ԧ

𝑑𝑡2= −𝑞𝐸ሬԦ0𝑒𝑖𝜔𝑡 − 𝑚𝜑

𝑑𝑟Ԧ

𝑑𝑡 , (3.2)


27

where 𝒎𝒆 and 𝒒 are the mass and the absolute charge of the electron, and 𝑬ሬሬԦ𝟎 defines the

amplitude of EM wave with angular frequency of 𝝎 . The second order differential

Equation (3.2), has a following general solution10:

Here, the imaginary part of the solution describes the phase shift of electron’s oscillation

with respect to the incident wave. It is apparent that, if the dissipative forces or in other

words the collision frequency is higher, the electron would have a delayed response

(damping) which is represented by the imaginary part in Equation (3.3). Next, to obtain

the polarization in the conductor generated by means of an external field, the total

contribution of electrons in the Drude model must be included. In this way, assuming that

the density of electrons is 𝒏, the scalar form of polarization ‖𝑷ሬሬԦ(𝒕)‖ induced by spatial

displacement ‖𝒓ሬԦ(𝒕)‖ will be:

Moreover, using the definition of electric displacement ‖𝑫ሬሬԦ(𝒕)‖:

the relative dielectric permittivity function (dispersion relation) of the conductor material

𝞮𝒓(𝝎) can be retrieved:

‖𝑟Ԧ(𝑡)‖ =𝑞‖𝐸ሬԦ0‖𝑒𝑖𝜔0𝑡

𝑚(𝜔2 + 𝑖𝜑𝜔) . (3.3)

‖𝑃ሬԦ(𝑡)‖ = −𝑛𝑞‖𝑟Ԧ(𝑡)‖ =−𝑛𝑞2

𝑚(𝜔2 + 𝑖𝜑𝜔) ‖𝐸ሬԦ0‖𝑒𝑖𝜔𝑡 . (3.4)

‖𝐷ሬሬԦ(𝑡)‖ = 𝜀‖𝐸ሬԦ0‖𝑒𝑖𝜔𝑡 = 𝜀0‖𝐸ሬԦ0‖𝑒𝑖𝜔𝑡 + ‖𝑃ሬԦ(𝑡)‖ , (3.5)

𝜀𝑟(𝜔) = 1 −

𝑛𝑞2

𝜀0𝑚

𝜔2 + 𝑖𝜑𝜔= 1 −

𝜔𝑝

𝜔2 + 𝑖𝜑𝜔 ,

(3.6)


28

where 𝞮𝟎 is the permittivity of free space and 𝝎𝒑 =𝒏𝒒𝟐

𝞮𝟎𝒎 is the plasma frequency10 of

electron cloud.

It should be highlighted that the final form of Equation (3.6) was acquired using

the Drude model of free electron gas. Considering the electronic bands of noble metals,

the Drude model requires some extensions in order to depict a more realistic behavior of

charge carriers. First of all, the internal polarization caused by positively charged metal

ions can be represented by a new dielectric constant 𝜺∞ instead of free space permittivity.

Additionally, the effect of electronic interband transitions can be classically interpreted

by including the internal bound electrons11. Thus, the second order differential equation

in this case would be:

where the 𝝎𝟎 term suggest angular frequency at which the interband transition occurs.

However, the final form of Equation (3.6) would be sufficient to have an illustrative

depiction of noble metal’s behavior under the impact of external EM wave.

3.2 Scattering on Spherical Particles

Now, having the classical interpretation and a quantitative model which

describes the behavior of noble metals under the act of external EM field, an idealized

example system can be used to illustrate the basic principle of the EM boosting

mechanism. Along with that, one of the fundamental excitation principle in plasmonics,

localized surface plasmons (LSP) will be introduced. Also, it will be shown that under

specific circumstances at the dielectric interface, a resonance condition and consequently

significantly higher EM field enhancement can be achieved. These core elements will be

𝑚𝑒

𝑑2𝑟Ԧ

𝑑𝑡2= −𝑞𝐸ሬԦ0𝑒𝑖𝜔𝑡 − 𝑚𝜑

𝑑𝑟Ԧ

𝑑𝑡− 𝑚𝜔0

2𝑟Ԧ , (3.7)


29

demonstrated using the quasi-static Mie model, the interaction of homogeneous spherical

particle with the external EM wave. In this way, the roughened electrode surface which

was referred in the introductory part of this chapter will be assumed to be spherical

particle or more precisely an arrangement of spherical particles12,13. Moreover, the quasi-

static approximation requires the diameter of spherical metallic particle to be negligible

as compared to wavelength of the EM radiation (𝞴 ≫ 𝒅)10. The electrodynamic Mie

problem is greatly simplified by quasi-static approach which allows to employ

electrostatic assumption for the model. Moreover, cylindrical coordinate system will be

utilized in order to provide clarity in the following discussions and analytical

formulations.

Figure 3.2 The schematics of general Mie problem. In the Mie model, illumination and as a

result scattering on a spherical particle with diameter 𝒅 and dielectric permittivity 𝞮(𝝎) placed

in non-absorbing environment (e.g. vacuum 𝞮𝟎 ) is studied. Due to spherical symmetry, an

analytical solution of the Mie scattering problem can be obtained.


30

The graphical representation of Mie scattering problem on a spherical particle is

shown in Figure 3.2. Due to the assumption 𝞴 ≫ 𝒅 , the electrodynamic problem is

reduced to the electrostatic Laplace equation which can be used to retrieve the electric

potential (𝑽) values at a given spatial position 𝒓 . In other words, the electric field

component of EM wave (𝑬𝟎𝒆𝒊(𝝎𝒕+𝒌ሬሬԦ⋅𝒓ሬԦ)) with small wavenumber 𝒌ሬሬԦ would imply slowly

developing field oscillation on a spherical particle. Thus, since the electric field

stimulation can be considered to be uniform on the particle, at a given time the

electrostatic condition of the system would be characterized using the following equation:

and correspondingly the electric field at a given point will be:

The general solution of differential Equation (3.8) is14:

Where the 𝑷𝒍(𝒄𝒐𝒔𝜽) term is the 𝒍’th order Legendre Polynomials and 𝑨𝒍 , 𝑩𝒍 are the

constants which can be determined from the boundary conditions. In order to obtain the

boundary conditions and subsequently the potential values inside 𝑽𝒊𝒏 and outside 𝑽𝒐𝒖𝒕

of the scattering object, a specific phase of time dependent EM field should be considered.

In this particular case, a stage where the electric field component is maximum (𝑬𝟎) will

be taken into account. In such manner, using the Gauss law, two boundary conditions for

electric displacement at the surface of particle 𝒓 =𝒅

𝟐 and electric field at 𝒓 → ∞ can be

formulated in the following way15:

𝛻 2𝑉 = 0 , (3.8)

𝐸ሬԦ = −𝛻ሬԦ 𝑉 . (3.9)

𝑉(𝑟, 𝜃) = ∑(𝐴𝑙𝑟𝑙

∞

𝑙=0

+ 𝐵𝑙𝑟−(𝑙+1) )𝑃𝑙(𝑐𝑜𝑠𝜃) , (3.10)


31

By applying the boundary conditions Equation (3.11) and (3.12), the potential values

inside and outside the spherical scattering particle according to Equation (3.10) will be:

The last term of Equation (3.14) can interpreted as spatial electric potential distribution

of a point dipole source10. Since the constrain for the quasi-static approximation is based

on 𝞴 ≫ 𝒅 condition, the interaction between EM wave and the spherical particle is

occurring in a rather uniform way. More precisely, taking into consideration the 𝞴 ≈ 𝒅

condition, the influence of field components on the scattering object would not be even

due to the spatial phase differences of the propagating EM wave. Likewise, in the quasi-

static approximation the phase variance is negligible when compared to the dimensions

of the particle. The particle can thus be treated as a point dipole source.

3.3 Localized Surface Plasmons

Despite of the fact that the proposed Mie model is neglecting the accurate nature

of light-spherical particle interaction, Equation (3.14) depicts to some extend the

dependence of spatial electric field distribution on geometrical parameter (diameter) and

optical properties of noble metals. The precision of the analytical approach can be

−𝜀 𝛻ሬԦ 𝑉𝑖𝑛|𝑟=

𝑑2

= −𝜀0 𝛻ሬԦ 𝑉𝑜𝑢𝑡|𝑟=

𝑑2

, (3.11)

−𝜵ሬሬԦ 𝑽𝒊𝒏|𝒓→∞

= − 𝜵ሬሬԦ 𝑽𝒐𝒖𝒕|𝒓→∞

. (3.12)

𝑉𝑖𝑛 = −3𝜀0

𝜀 + 2𝜀0 𝐸0𝑟 𝑐𝑜𝑠𝜃 , (3.13)

𝑽𝒐𝒖𝒕 = −𝑬𝟎𝒓 𝒄𝒐𝒔𝜽 + 𝑬𝟎 (𝒅

𝟐)

𝟑 𝒄𝒐𝒔𝜽

𝒓𝟐

𝜺 − 𝜺𝟎

𝜺 + 𝟐𝜺𝟎 . (3.14)


32

expanded by introducing further adjustments. For instance, using modified long

wavelength approximation (MLWA)16 in the Mie model, analytical solutions with better

correlation as compared to experimental observation can be achieved17. Additionally, the

Drude approach, which describes the optical properties (dielectric permittivity function)

Equation (3.6) of conducting materials can be exploited in the Mie model to demonstrate

the behavior of spherical noble metal particles under the impact of EM wave. Prior to

that, it should be noted that the distinct solutions obtained from the quasi-static Mie model

is corresponding to one particular phase of EM oscillation. Hence, to get the full picture

of light-metallic sphere interaction, a time dependent study of the Mie system should be

conducted. In this way, considering only the quasi-static model, the linear field

dependence in Equation (3.13) and (3.14) is indicating that the particle can be treated as

an oscillating point dipole under the effect of propagating EM wave. This time dependent

action is demonstrated in Figure 3.3.

The physical explanation of such time dependent variation in polarization of the

spherical conducting particle under the impact of EM field (Figure 3.3) is accounted for

spatial shifts of local unbound electrons. Thus, the phenomenon which proposes

oscillatory motion of local charge carriers in sub-wavelength conducting objects

facilitated by means of EM perturbation is in literature referred to as LSP excitation10,18.

Further, from the Drude representation of conductors Equation (3.6), it is evident that the

response of a conducting material is dependent on the wavelength of incident EM

radiation (from the dispersion relation). Hence, for a particular plasmonic material (e.g.

Ag), the wavelength of external illumination would influence the degree of LSP

excitation and as a consequence the amplitude of the dipole oscillation. This is originating

from the fact that, any kind of oscillatory system (e.g. Drude model) implies a natural

mode (resonance frequency) at which the energy transfer from the source (EM radiation)

to the system (electron cloud motion) is maximized19. Moreover, one of the significant


33

outcomes of the LSP excitation is the light scattering phenomenon which occurs through

oscillatory motion of “dipole” source (Rayleigh scattering). In fact, the dipole radiation

is angle dependent and it implies that the radiated light would have preserved wavelength

compared to the incident illumination. In the following, LSP resonance and resulting

boosting of the near-field EM radiation on plasmonic particles will be presented. One

should consider that the analytical Mie model is only suitable for the spherical particles.

On top of that, even considering the spherical materials, the analytical approach is not

suitable for studying arrangement of interacting particles. Alternatively, additional

computational methods e.g. finite-difference time-domain (FDTD) technique can be

utilized to imitate the field response of a conducting material with a custom morphology

Figure 3.3 Electron cloud oscillation in spherical particles under the impact of external EM

wave. The unbound electron cloud and positively charged ions of spherical metallic particle are

illustrated using green and grey colors respectively. Here, time dependent shift of electron clouds

under the effect of propagating EM wave and as a result polarization of the particle are

schematically represented.


34

(e.g. spherical dimer structures). Briefly, the FDTD method employs the discretized form

of the electrodynamic Maxwell equations20:

to presume time dependent response of the simulated environment. Here, 𝑫ሬሬԦ stands for

electric displacement, 𝝆𝒇 unbound free charge density, 𝑩ሬሬԦ magnetic field, 𝑯ሬሬሬԦ auxiliary

magnetic field and 𝑱𝒇ሬሬሬԦ free current density respectively.

The degree of LSP excitation, scattering cross-section for spherical Au dimer

structures calculated through FDTD computational method, is depicted in Figure 3.4. The

resonance condition can be clearly visualized in Figure 3.4c, d where a peak in scattering

cross-section or maximized energy transfer between incident illumination and the

plasmonic particles is achieved at ~530 nm. Alternatively, the case shown in Figure 3.4a,

b, proposes LSP resonance peak shifts with changing gap sizes. Here, another practical

use of computational methods for LSP applications is demonstrated. Unlike the Mie

model where the analytical model is achieved through tedious derivations (Equations

(3.13) and (3.14)), FDTD modelling provides opportunity to imitate the interaction

between distant plasmonic objects. In this case, the electron cloud oscillation in both of

the spherical Au particles throughout the EM wave perturbation process are

𝛻ሬԦ · 𝐷ሬሬԦ = 𝜌𝑓 , (3.15)

𝜵ሬሬԦ · 𝑩ሬሬԦ = 𝟎 , (3.16)

𝜵ሬሬԦ×𝑬ሬሬԦ = −𝝏𝑩ሬሬԦ

𝝏𝒕 , (3.17)

𝜵ሬሬԦ ×𝑯ሬሬሬԦ = 𝑱𝒇ሬሬሬԦ +

𝝏𝑫ሬሬԦ

𝝏𝒕 , (3.18)


35

electrostatically influencing (“dipole-dipole” interaction) each other. As a consequence,

the natural LSP mode of the spherical Au particles is shifted21.

Another essential point for understanding the EM boosting mechanism of

surface-enhanced RS is related to the near-field scattering phenomenon of LSP excitation.

A typical illustration for the localized scattering intensity at the LSP resonance condition

is shown in Figure 3.5. Here, the spatial distribution of the second power of the relative

electric field (with respect to incident field) or the relative scattering intensity for 10 nm

dimer gap size and geometrical condition summarized in Figure 3.4a is presented. The

resonance condition from Figure 3.4b suggest incident illumination at 563 nm wavelength

Figure 3.4 Scattering cross-section of spherical dimer structures with different gap

distances. (a, b) Gab distance dependent cross-section calculations of spherical dimer structures.

The electric field direction of incident illumination (polarization) is towards the dimer orientation.

(c, d) Similarly, the calculated scattering cross- section of dimer s with perpendicular

arrangement. Adapted from reference23.


36

at which the FDTD simulations in Figure 3.5 were performed. Further, from Figure 3.5,

the “hot-spots” or the spot with highest scattering intensities can be easily determined. A

parallel depiction can be made with optical lenses which enables the incident light beam

to be focused in a comparably narrow spot. Similarly, the plasmonic particles, in this case

Au dimer spheres are focusing the incident coherent and monochromatic beam in

nanoscale “hot-spots”. It is obvious that, depending on the morphology of the plasmonic

material and in broader sense, the arrangement of multiple plasmonic particles,

determines the LSP modes and consequently the localized scattering intensities. In this

way, plasmonic engineering of subwavelength conducting particles can be utilized for

manipulation of incident radiation in nanoscale dimensions.

Figure 3.5 LSP excitation and scattering on the spherical Au dimer structures. The incident

EM illumination with 563 nm wavelength (resonance condition) results in LSP excitation in

spherical dimer structures. Subsequently, the induced scattering effect is summarized in terms of

spatial square of the relative electric field (intensity) arrangement in logarithmic scale. Adapted

from reference23.


37

3.4 EM Enhancement Mechanism

The enhancement phenomenon of RS is related to near-field properties of

plasmonic noble metal surfaces under influence of external illumination. For the case with

spherical dimer structures, it was shown in Figure 3.5 that the field response close to the

surface (especially at the gap) is substantially increased. Assuming a Raman active

molecule placed in this EM “hot-spot”, an evident enhancement of RS due to localized

intensity of exciting illumination can be interpreted (see Equation (3.1) for RS intensity).

Indeed, to get a full picture of the system, the electro-dynamic system of plasmonic

surface and the molecule of interest (Raman probe) should be studied. The RS boosting

mechanism which is a two-step process, is dependent on synergistic EM interaction of a

plasmonic rough surface and molecule system with the incident radiation22. The electron

band interaction between the surface of the noble metal and the Raman active analyte can

give rise to another boosting mechanism (chemical enhancement). Because, the chemical

boosting mechanism is highly dependent on the type of molecule and the interacting

plasmonic material, the effect will be neglected. Moreover, as reported in various studies,

in most of the cases the main contribution to the enhancement factor (EF) is accounted

for by the EM mechanism. Typically, chemical EF is ranging from 10 to 102 while the

EM mechanism can yield EF up to 108 (depending on morphological and material

properties)7.

In Figure 3.6, the general concept of a two-step EM enhancement mechanism of

RS on plasmonic surfaces is presented. The first contributing boosting mechanism is

obtained through LSP excitation of metallic particles. For simplicity, a spherical metal

dimer structures will be given as an example. In Section 3.3, the localized enhancement

(as compared to incident radiation, see Figure 3.5) of EM field as a result of LSP

excitation was given. The LSP excitation and the following scattering effect occurs in


38

agreement with the Rayleigh model. As a consequence, the locally scattered light has the

same angular frequency 𝝎𝟎 as the incident radiation4. Thus, by realizing Raman

spectroscopy measurement of Raman active molecules, e.g. melamine on spherical Au

dimers, according to Equation (3.1) the EF of Raman signal can be expressed as:

where 𝑬𝒍𝒐𝒄(𝝎𝟎) is the locally enhanced electric field on plasmonic particle and 𝑬𝒊𝒏𝒄(𝝎𝟎)

is the magnitude of the electric field component of the incident EM wave. In this way, the

𝐺1 ≈|𝐸𝑙𝑜𝑐(𝜔0)|2

|𝐸𝑖𝑛𝑐(𝜔0)|2 , (3.19)

Figure 3.6 Two-step EM enhancement mechanism of RS. The first stage of EM boosting

principle is achieved through near-field intensification of incident radiation with angular

frequency 𝝎𝟎 on the surface of plasmonic material. Further, the Raman active molecule located

in the plasmonic “hot-spot” would generate enhanced RS effect. Lastly, the second step proposes

re-radiation of RS through plasmonic particles.


39

boosted RS signal with angular frequency 𝝎𝑹𝑺 is obtained (see Equation (2.13)).

Additionally, it is clearly illustrated for the Au dimer case (Figure 3.4) that the optical

properties of noble metals inhere a broad range of natural LSP excitation modes.

Subsequently, the second stage of EM mechanism suggests boosting of the RS though re-

radiation on plasmon active surfaces. In a similar way, the second input to the

enhancement can be given as:

Here, 𝑬𝑹𝑺(𝝎𝑹𝑺) is the amplitude of the electric field component of RS and 𝑬𝒍𝒐𝒄(𝝎𝑹𝑺) is

the locally enhanced field due to re-radiation effect on plasmonic structures.

It should be emphasized that the re-radiation of RS is accomplished for EM wave

with angular frequency of 𝝎𝑹𝑺 . In fact, the RS contains spectroscopically rich

information about the vibrational modes of the molecule through the shifts in angular

frequency 𝝎𝟎 of incident illumination. Hence, in reality the 𝝎𝑹𝑺 value represents a set of

angular frequencies (see Figure 2.5). Now, taking into account that the LSP excitation is

actually wavelength dependent (Figure 3.4), the second step of the EM boosting process

would result in diversified EF for a distinct 𝝎𝑹𝑺 value. This means that for a Raman

active molecule, the original Raman spectrum would be distorted in the surface-

enhancement process. The change in fingerprint spectrum would be determined by the

LSP behaviour of the plasmonic surface.

To conclude, a complete EF for a particular vibrational mode obtained from both

of the contributions is given as:

𝐺2 ≈|𝐸𝑙𝑜𝑐(𝜔𝑅𝑆)|2

|𝐸𝑅𝑆(𝜔𝑅𝑆)|2 . (3.20)

𝐸𝐹 =𝐼𝑆𝐸𝑅𝑆

𝐼𝑅𝑎𝑚𝑎𝑛≈ 𝐺1𝐺2 =

|𝐸𝑙𝑜𝑐(𝜔0)|2

|𝐸𝑖𝑛𝑐(𝜔0)|2 |𝐸𝑙𝑜𝑐(𝜔𝑅𝑆)|2

|𝐸𝑅𝑆(𝜔𝑅𝑆)|2 . (3.21)


40

This formulation describes the EF achieved on a single molecule. For practical

applications, it is more accurate to consider the number of probed molecules and local

field distribution3,4,22. Nevertheless, Equation (3.21) provides an illustrative

representation of the EM boosting mechanism of the surface-enhanced RS phenomenon.

References

(1) Van Duyne, R. P. Chem. Biochem. Appl. lasers 1979, 4, 101.


163–166.

(3) Stiles, P. L.; Dieringer, J. A.; Shah, N. C.; Van Duyne, R. P. Annu. Rev. Anal.

Chem. 2008, 1 (1), 601–626.

(4) Ding, S.-Y.; You, E.-M.; Tian, Z.-Q.; Moskovits, M. Chem. Soc. Rev. 2017, 46

(13), 4042–4076.

(5) Otto, A. Appl. Surf. Sci. 1980, 6 (3–4), 309–355.

(6) Campion, A.; Ivanecky, J. E.; Child, C. M.; Foster, M. J. Am. Chem. Soc. 1995,

117 (47), 11807–11808.

(7) Kiefer, W. Surface enhanced Raman spectroscopy: analytical, biophysical and

life science applications; John Wiley & Sons, 2011.

(8) Wooten, F. Optical properties of solids; Academic press, 2013.

(9) Etchegoin, P. G.; Le Ru, E. C.; Meyer, M. J. Chem. Phys. 2006, 125 (16),

164705.

(10) Maier, S. A. Plasmonics: fundamentals and applications; Springer Science &

Business Media, 2007.

(11) Schulz, L. G. Adv. Phys. 1957, 6 (21), 102–144.

(12) Mishchenko, M. I.; Hovenier, J. W.; Travis, L. D. Light scattering by

nonspherical particles: theory, measurements, and applications; Academic press, 1999.

(13) Creighton, J. A.; Blatchford, C. G.; Albrecht, M. G. J. Chem. Soc. Faraday

Trans. 2 1979, 75, 790.

(14) Jackson, J. D. Electrodynamics; Wiley Online Library, 1975.

(15) Griffiths, D. J. Introduction to Electrodynamics; Prentice Hall, 1999.


41

(16) Geddes, C. D. Reviews in Plasmonics 2016; Springer, 2017.

(17) Wriedt, T. Springer, Berlin, Heidelberg, 2012; pp 53–71.

(18) Hutter, E.; Fendler, J. H. Adv. Mater. 2004, 16 (19), 1685–1706.

(19) Mayer, K. M.; Hafner, J. H. Chem. Rev. 2011, 111 (6), 3828–3857.

(20) Taflove, A.; Hagness, S. C. Computational electrodynamics: the finite-

difference time-domain method; Artech house, 2005.

(21) Hong, Y.; Huh, Y.-M.; Yoon, D. S.; Yang, J. J. Nanomater. 2012, 2012, 1–13.

(22) Ding, S.-Y.; Yi, J.; Li, J.-F.; Ren, B.; Wu, D.-Y.; Panneerselvam, R.; Tian, Z.-

Q. Nat. Rev. Mater. 2016, 1 (6), 16021.

(23) Le, K. Q.; Alù, A.; Bai, J. J. Appl. Phys. 2015, 117 (2), 23118.


42

Chapter 4: Plasmonic NP Structures

43

4 PLASMONIC NP STRUCTURES

Since its discovery in 1970’s1, surface-enhanced RS gradually attracted interest

as a spectroscopic method for various practical applications. First of all, due to the

quenching effect of fluorescent molecules on conducting surfaces, the commonly

encountered background issue in Raman spectroscopy was resolved using surface-

enhanced Raman spectroscopy (SERS) technique2. Furthermore, SERS active surfaces

(plasmonic structures) has crucial impact when it is employed for molecular sensing.

With the advances in nano-fabrication and chemical synthesis methodologies, a broad

variety of plasmonic structures including colloidal solution of nanoparticles, thin film of

nanoparticle aggregates and highly ordered nanostructures obtained via

lithography\imprint technology developed3–5. However, depending on the practical

utilization of SERS there is an emerging need to evaluate the performance of active

surfaces.

The sensing platform, proposed in this PhD thesis, should facilitate qualitative

as well as quantitative analysis of analytes in a sample suspension. The SERS technique

is highly sensitive, even single molecule detection has been reported6. This is achieved

via SERS active surfaces with tremendously high localized EF. However, in most of the

cases, due to errors in fabrication and as a result local variations in geometry of plasmonic

structures, localized high EF are not uniform. This is originating from the near-field

properties of LSP excitation which becomes uncontrolled for e.g. aggregated particles

with sub-nm gap distances7. Alternatively, electron-beam lithography processes can be

exploited to achieve highly ordered SERS active substrates8,9. But, in that case, one


44

should consider the ease and cost of the fabrication procedure. For quantitative

assessment of analyzed sample, quality parameters for the active SERS surfaces can be

formulated. Primarily, a SERS substrate should exhibit EF above 105 to achieve the

desired sensitivity7. Further, quantitative study in SERS requires statistical treatment of

the data, hence the EF should be uniform in macroscale with variations of less than 20%.

Next, the fabrication process should be reproducible, easy to perform and cheap. Lastly,

the plasmonic surface should be clean (free of contamination e.g. arising from fabrication)

and provide low interfering SERS background10.

In this chapter, the sensing tool, metal coated nanopillar SERS substrates which

was employed in this PhD work will be presented. Also, complementary utilization of NP

SERS substrates will be discussed.

4.1 Fabrication of Metal Coated Si NP Structures

The fabrication procedure for plasmonic NP structures11 is based on a modified

black Si etching process. In our research group, over the past years, multiple NP based

SERS substrates have been developed. They include optimizations of LSP excitation

modes12 and density, height of NPs for uniformity of EF12–14, flexibility in terms of base

material (Si, SiO2, polymeric materials) which eventually provides versatility of NP based

SERS substrates for various practical sensing implementations15–18. More importantly,

the fabrication technique is maskless. This means that throughout the process, costly

lithography steps are not implemented. In addition to that, the overall fabrication

procedure for metal coated Si NP surfaces yield batch to batch reproducibility in terms of

plasmonic properties11,13.

In general, the fabrication of Si NP SERS substrates consists of three main steps.

The overview of the procedure is shown in Figure 4.1. First, NP structures (~50-80 nm


45

wide) are obtained on a Si wafer using maskless reactive ion etching (RIE) technique with

SF6\O2 process gasses (Figure 4.1a). The RIE is a very delicate method, and by carefully

Figure 4.1 Summary for fabrication process of Si base metal coated NP structures. (a)

Maskless RIE of Si surface using SF6\O2 process gasses. (b) Upon completion, NPs on Si surface

are obtained and the contaminants from RIE step are removed by O2 plasma cleaning procedure.

(c) Metallization step. Using one of the traditional deposition techniques e.g. electron-beam

evaporation, the desired thickness of metal is obtained on Si NPs. (d) SEM image Si NP structures

with Au caps achieved via electron-beam evaporation method.


46

tuning the process parameters, various micro and nanostructures can be obtained. Thus,

the parameters of RIE can be adjusted to achieve different pillar heights and densities11.

Usually, the etching time is linearly correlated with the NP’s height and by manipulating

the pressure in the etching chamber, the density of NPs can be altered. A SERS substrate

should demonstrate low background signal. Therefore, the RIE of Si is followed by an O2

plasma cleaning of the surface to eliminate interference of sulphur-fluoride based etching

by-products14 (Figure 4.1b). Lastly, the plasmonic property of Si NPs is obtained by a

metallization step (Figure 4.1c). Using a conventional deposition technique, for instance

the electron-beam evaporation technique, noble metals (e.g. Au, Ag or Al) with a desired

thickness are applied on the NP surface. The final morphology of metal coated (in this

case 160 nm of Au via electron-beam method) Si NPs with ~400 nm height is depicted in

Figure 4.1d. The image was acquired via scanning electron microscopy (SEM) technique.

Different means of metal deposition would develop variations in morphology of

the metal on the NPs. For instance, using magnetron sputtering of Au on NP structures

results in uniform coating of the metal while the electron-beam evaporation technique

facilitates formation of “caps” on top side of NPs11. Consequently, the LSP excitation

modes and SERS performance are affected by the alterations in the metal coating. In the

following deliberations, the NP structures achieved by directional e-beam evaporation

technique, will be considered.

4.2 Leaning Effect of NP Structures

Normally, the applications of fluid samples on SERS substrates are performed

by placing a droplet on the surface. According to wettability properties of the liquid

towards the nanostructured surface, the sample would locally pin on the applied region

or spread across the dry regions and finally evaporate. However, for the case with the NP

structures, by tuning the mechanical properties (e.g. stiffness) through modification of the


47

NPs geometry, a clustering or self-assembly of NPs with the nearest neighbors can be

achieved (compare Figure 4.2a the nonwetted region and Figure 4.2b the area exposed to

the liquid). In a straightforward way, this can be promoted by tuning the height of NPs

via RIE time. This clustering or leaning effect of NPs is greatly dependent on surface

tension of the applied liquid sample. Thus, throughout the liquid sample evaporation

process, due to capillary action on the nanotextured surface11, the NPs at the localized

spots are pulled towards each other. As per observations reported in various studies with

analogous NP SERS substrates, water, ethanol, acetone and other commonly used

solvents are facilitating clustering effect46,49. Such mechanical self-clustering process

would eventually result in formation of sub-nm gap sizes and as a consequence highly

localized field enhancements, “hot-spots” as it was shown for the spherical dimer case

(Figure 3.4). In Section 4.3, the optical properties of NP structures will be elaborated.

Figure 4.2 Clustering of NPs induced by capillary forces. (a) Top view SEM image of dry

region of the SERS substrate, the leaning effect of NPs is not observed. (b) Wetted region, the

SEM image was captured right after the evaporation of fluid. The capillary forces during the

evaporation process, promoting clustering of the structures. The Si NPs were 400 nm in height

and deposited Au thickness was 160 nm.


48

4.3 Plasmonic Properties of NP Substrates

In Section 3.3, it was shown that the LSP excitation modes as well as localized

field enhancement through the scattering effect on conducting particles are significantly

dependent on the material, geometric properties and arrangement of distinct particles.

Similarly, plasmonic properties of NP structures are considerably determined by the

cavity mode, a narrow range LSP excitation originating from the metal cap-Si NP

interface12. By modifying the Si NP diameter and correspondingly changing the gap size

of the metal cap, the cavity mode of LSP excitation can be tuned. Nevertheless, due to

the errors in RIE process, in reality the NP diameter varies in microscopic regions. For

instance, using dark-field instrument which allows to probe scattering on large areas, the

distinct plasmonic modes are overlapped resulting in a broader LSP response. On top of

that, the leaning effect promotes generation of hybridized mode which is achieved

through interaction of LSP modes on diverse NPs12. An example for that is demonstrated

in Figure 4.3a where experimental dark-field study on Ag coated Si NP structures were

conducted. Such broad range LSP resonance response is very attractive for SERS based

sensing systems. Depending on the medium, to avoid interferences like fluorescence

emission, the excitation laser wavelength of Raman spectrometer can be selected within

the LSP resonance band. In Figure 4.3b, simulated electric field enhancement distribution

on dimer structure of 40 nm wide Si NP and 310 nm in height, 124 nm wide Ag caps

excited using various illumination wavelengths are shown. Taking into account the EM

enhancement mechanism elaborated in Section 3.4, Figure 4.3b in an illustrative way

demonstrates the spatial arrangement of field EF and consequently different LSP

resonance modes suitable for molecular probing through SERS. Alternatively, the broad

range LSP resonance bands can be tuned by substituting the Ag cap with Au or Al.


49

The essential criteria for SERS performance relies on EF and its macroscale

uniformity. The investigation on Ag capped NP substrates showed that under optimal

conditions, EF ~1.3 × 108 can be potentially reached. Furthermore, the signal

reproducibility or uniformity of EF across a macroscale region (on a 5×5 mm2 chip) ~14%

was reported14. In summary, with respect to all quality factors and as compared to the

commercially available and reported SERS active surfaces7, NP substrates exhibit

exceptionally good performance as a SERS substrate. Additionally, for in-field

applications, the sustainability or shelf-life of the sensor is important. Because of the

oxidation effect, Ag coating of NPs can be substituted with Au.

Figure 4.3 LSP excitation and simulated EF distribution on Ag coated NP dimers. (a)

Measured dark-field scattering on NP surface before (left) and after the leaning (right) took place.

The different colours correspond to distinct Ag deposition thicknesses. The circle, star and

asterisks are representing the hybridized, particle and cavity LSP modes respectively. (b)

Simulated relative field distribution of NP dimers at various excitation wavelengths. Adapted

from reference12.


50

4.4 Sensing Applications using SERS Substrates

SERS is a surface selective sensing technique, and a well-engineered plasmonic

substrates can provide tremendous enhancement of RS arising from analyte molecules.

However, from a practical point of view, utilization of SERS active substrates in most of

the cases, especially for real-life fluid samples, is problematic. Commonly accepted issues

to be addressed are; (i) chemically rich environment of the sample20, (ii) surface

selectivity of the sensing platform and lastly21, (iii) the sample application method onto

the substrate surface22,23. The first two points are mostly determining the sensitivity and

detection limit of a SERS measurement and the last one is correlated with the

implementation of a SERS based sensor for molecular quantification purposes. In order

to outline the possible obstacles arising from these issues, a specific case will be given as

an example.

For instance, a straightforward implementation of SERS technique (applying

sample droplet on active surface) for detection of creatinine in human blood serum can

be prohibited by macromolecular enzymes and proteins present in the sample24. It should

be emphasized that the SERS technique is suitable only for molecular systems. The

proteins, as an example of macromolecules, would due to their complex molecular

structure appear as background in SERS measurements. Moreover, such structures are

clogging the plasmon active surfaces, completely blocking the RS from analyte

molecules. Thus, the macromolecular clusters are not desired for SERS based platforms.

Moreover, upon sample application on the substrate surface, there is a competing

molecular “binding” process onto the SERS active region. The arrangement of analyte

molecules will be mostly determined by their surface affinities towards the noble metal

(e.g. Au). The surface-enhanced RS phenomenon due to the near-field property of

plasmonic structures occurs selectively in proximate regions. Hence, RS of analytes,


51

which are tightly allocated on a SERS active region, would be preferably enhanced more

as compared to molecules with less affinity. In most of the recent SERS based studies

with complex media, surface functionalization of the active substrate region was proposed

to target specific analytes and to overcome the problem origination from the molecular

affinity25–28. Of course, such chemical engineering of the surface would eventually

increase the sensitivity of the spectroscopic technique. Nevertheless, much more sensitive

methods e.g. lateral flow based assays can be developed for detection of specific target

analyte in complex mixture29. The strength of SERS stands in its specificity for molecular

detection. A broad variety of molecules in a complex sample can be identified using a

fingerprint spectrum. Hence, the SERS based platforms should target identification and

quantification of broad range analyte molecules in complex sample mixture and affinity

based issues requires facile and robust sample handling-treatment approaches.

Lastly, the conventional quantification using SERS substrates is achieved

through statistical analysis of the data obtained on macroscopic regions (SERS

mapping)17. The surface-enhanced RS intensity is highly determined by the localized

concentration of Raman active molecules in the plasmonic hot-spots. Thus, together with

SERS EF uniformity, experimentally consistent spectral signal responses are requiring

uniform analyte distribution. The traditional sample droplet application onto the SERS

substrate would result in irregular molecular dispersion in macroscale regions known as

coffee ring effect30. The issue, as reported in multiple studies can be solved via combining

SERS based sensing with microfluidics platform. The microfluidics integration allows to

perform highly controlled liquid handling operations in an enclosed system eliminating

the external influences (e.g. uncontrolled evaporation process)31. On this account, a SERS

integrated centrifugal microfluidics platform, presented Chapter 0, was developed during

the PhD study. An additional obstacle is the SERS signal response with respect to

molecular concentration. The quantification using the characteristic non-linear sigmoidal


52

response named “Langmuir adsorption curve” is challenging for extremely high and low

analyte concentrations. In Chapter 0 and 0, the experimental findings addressing above

mentioned issues are further elaborated.

4.5 Electrochemical Sensing

Electrochemical (EC) based sensing platforms are dealing with chemical

reactions which occur at the electrode-sample interface. The degree of chemical

interaction at the interface is measured through the electrical quantities such as current,

potential or capacitance32. Modern EC based platforms involve a large portfolio of

techniques for analyzing chemical reactions in complex matrices. The commonly

implemented ones are categorized under the controlled-potential or potentiostatic

methods which are carried out on three electrode systems (working, counter and reference

electrodes). These techniques are studying the dynamic charge transfers at the liquid

sample-electrode interface32. In this case, an electric potential is applied between the

counter and working electrodes to induce an electron transfer. And the absolute electric

potential value at the distinct electrodes is calibrated through a reference electrode. An

arising e.g. electrical current can be measured to study the chemical reactions in the

sample solution. The overall goal in EC based platforms is to monitor the concentrations

of molecular species via chemical reactions or more specifically the redox process of the

analytes (Faradaic process)33:

at the electrode-sample interface. Here, 𝑶 and 𝑹 are the oxidized and reduced

compounds, the 𝑵 represents the number of electrons (𝒆−) transferred to the EC system

during the process. The redox process proposes a characteristic potential value 𝑽𝟎 at

which the reaction becomes possible. Thus, theoretically the concentration ratio between

𝑂 + 𝑁𝑒− 𝑅 , (4.1)


53

oxidized 𝑪𝑶 and reduced 𝑪𝑹 components can be estimated through the applied potential

𝑽 (Nernst equation)32:

where 𝑹 stands for universal gas constant (8.314 J×K-1×mol-1), 𝑻 temperature of the

system and 𝑭 is the Faraday constant (69.487 C). The resultant current 𝑰 (Faraday

current) which provides information about rate of the redox process is subsequently

measured for further EC based characterization of the sample solution. On this account,

several techniques such as cyclic voltammetry (CV) or amperometry can be individually

deployed to extract information about the quantitative composition of molecular species

in the sample solution. CV which implies dynamic study of current in the electrode system

versus varying electric potential is the most straightforward and basic method to obtain

qualitative assessment of the redox reaction. In addition to that, the quantification of the

analyte can be accomplished via the amperometric technique where current response over

time is studied at applied constant electric potential.

Instead of a tedious statistical SERS data analysis, as an alternative

complementary EC techniques can be integrated with SERS. Thus, using the molecular

specificity of SERS for qualitative assessment of the sample medium and the accuracy of

EC based measurements for analyte quantification, a synergistical combination of two

platforms can be achieved. A proof-of-concept EC-SERS approach is described in

Chapter 0.

References


163–166.

𝑉 = 𝑉0 +𝑅𝑇

𝑁𝐹 𝑙𝑛 (

𝐶𝑂

𝐶𝑅) , (4.2)


54


(3) Mosier-Boss, P. Nanomaterials 2017, 7 (6), 142.

(4) Sharma, B.; Frontiera, R. R.; Henry, A.-I.; Ringe, E.; Van Duyne, R. P. Mater.

Today 2012, 15 (1–2), 16–25.

(5) Fan, M.; Andrade, G. F. S.; Brolo, A. G. Anal. Chim. Acta 2011, 693 (1–2), 7–

25.

(6) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.;

Feld, M. S. Phys. Rev. Lett. 1997, 78 (9), 1667–1670.

(7) Prochazka, M. Surface-Enhanced Raman Spectroscopy: Bioanalytical,

Biomolecular and Medical Applications; Biological and Medical Physics, Biomedical

Engineering; Springer International Publishing, 2015.

(8) Hu, M.; Ou, F. S.; Wu, W.; Naumov, I.; Li, X.; Bratkovsky, A. M.; Williams, R.

S.; Li, Z. J. Am. Chem. Soc. 2010, 132 (37), 12820–12822.

(9) Abu Hatab, N. A.; Oran, J. M.; Sepaniak, M. J. ACS Nano 2008, 2 (2), 377–385.

(10) Sharma, B.; Fernanda Cardinal, M.; Kleinman, S. L.; Greeneltch, N. G.;

Frontiera, R. R.; Blaber, M. G.; Schatz, G. C.; Van Duyne, R. P. MRS Bull. 2013, 38 (8),

615–624.

(11) Schmidt, M. S.; Hübner, J.; Boisen, A. Adv. Mater. 2012, 24 (10), 11–18.

(12) Wu, K.; Rindzevicius, T.; Schmidt, M. S.; Mogensen, K. B.; Xiao, S.; Boisen,

A. Opt. Express 2015, 23 (10), 12965–12978.

(13) Wu, K.; Rindzevicius, T.; Schmidt, M. S.; Thilsted, A. H.; Boisen, A. J. Raman

Spectrosc. 2017, 48 (12), 1808–1818.

(14) Wu, K.; Rindzevicius, T.; Schmidt, M. S.; Mogensen, K. B.; Hakonen, A.;

Boisen, A. J. Phys. Chem. C 2015, 119 (4), 2053–2062.

(15) Morelli, L.; Andreasen, S. Z.; Jendresen, C. B.; Nielsen, A. T.; Emnéus, J.; Zór,

K.; Boisen, A. Analyst 2017, 142 (23), 4553–4559.

(16) Morelli, L.; Zór, K.; Jendresen, C. B.; Rindzevicius, T.; Schmidt, M. S.; Nielsen,

A. T.; Boisen, A. Anal. Chem. 2017, 89 (7), 3981–3987.

(17) Palla, M.; Bosco, F. G.; Yang, J.; Rindzevicius, T.; Alstrom, T. S.; Schmidt, M.

S.; Lin, Q.; Ju, J.; Boisen, A. RSC Adv. 2015, 5 (104), 85845–85853.

(18) Yang, J.; Palla, M.; Bosco, F. G.; Rindzevicius, T.; Alstrøm, T. S.; Schmidt, M.

S.; Boisen, A.; Ju, J.; Lin, Q. ACS Nano 2013, 7 (6), 5350–5359.


55

(19) Durucan, O.; Rindzevicius, T.; Schmidt, M. S.; Matteucci, M.; Boisen, A. ACS

Sensors 2017, 2 (10), 1400–1404.

(20) Sun, F.; Hung, H.; Sinclair, A.; Zhang, P.; Bai, T.; Galvan, D. D.; Jain, P.; Li,

B.; Jiang, S.; Yu, Q. Nat. Commun. 2016, 7, 13437.

(21) Bell, S. E. J.; Sirimuthu, N. M. S. J. Phys. Chem. A 2005, 109 (33), 7405–7410.

(22) Lee, M.; Lee, K.; Kim, K. H.; Oh, K. W.; Choo, J. Lab Chip 2012, 12 (19), 3720.

(23) Strehle, K. R.; Cialla, D.; Rösch, P.; Henkel, T.; Köhler, M.; Popp, J. Anal.

Chem. 2007, 79 (4), 1542–1547.

(24) Stosch, R.; Henrion, A.; Schiel, D.; Güttler, B. Anal. Chem. 2005, 77 (22), 7386–

7392.

(25) Etchegoin, P. G.; Le Ru, E. C. Phys. Chem. Chem. Phys. 2008, 10 (40), 6079.

(26) Liu, T.-Y.; Tsai, K.-T.; Wang, H.-H.; Chen, Y.-H.; Chen, Y.-H.; Chao, Y.-C.;

Chang, H.-H.; Lin, C.-H.; Wang, J.-K.; Wang, Y.-L. Nat. Commun. 2011, 2, 538.

(27) Dinish, U. S.; Yaw, F. C.; Agarwal, A.; Olivo, M. Biosens. Bioelectron. 2011,

26 (5), 1987–1992.

(28) Pekdemir, M. E.; Ertürkan, D.; Külah, H.; Boyacı, İ. H.; Özgen, C.; Tamer, U.

Analyst 2012, 137 (20), 4834.

(29) Koczula, K. M.; Gallotta, A. Essays Biochem. 2016, 60 (1), 111–120.

(30) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T.

A. Nature 1997, 389 (6653), 827–829.

(31) Huang, L.; Bian, S.; Cheng, Y.; Shi, G.; Liu, P.; Ye, X.; Wang, W.

Biomicrofluidics 2017, 11 (1), 11501.

(32) Wang, J. Analytical electrochemistry; John Wiley & Sons, 2006.

(33) Heineman, W. R.; Kissinger, P. T. Anal. Chem. 1980, 52 (5), 138–151.


56

Chapter 5: Centrifugal Microfluidics

57

5 CENTRIFUGAL MICROFLUIDICS

Miniaturized microfluidic systems have attracted great interest for various point-

of-care, environmental and food safety applications. The advantage of microfluidics

based platforms over the existing diagnostic techniques are their unique capabilities in

liquid sample manipulations, possible in-field employment and sensor integration. Thus,

complicated laboratory fluid handling steps such as metering, mixing, valving and sample

transport can be efficiently accomplished in an automated way without any human

interference1,2. In a conventional microfluidics device, the fluid is driven through

dynamic control of the pressure in an enclosed chip. This is primarily achieved by active

utilization of external pumps and in this way, the hydrodynamic properties such as sample

flow rate are precisely regulated3,4. However, as the practical requirements for sample

pre-treatment expand, the design and fabrication processes of the multifunctional

microfluidic device are becoming tedious and difficult to realize. Moreover, most of the

point-of-care applications are necessitating multiplex analysis of the sample. As a result,

the final device may consist of a large number of external pumps, tubings connecting to

microfluidic chip and sophisticated structures for fluid handling5,6.

These practical complications can be partially resolved by centrifugal

microfluidics7,8. In contrast to conventional microfluidics, the centrifugal platform does

not require any external pumping devices and tubings. The fluid management is realized

through inertial forces acting on fluids which are promoted by rotation of the microfluidic

disc platform7. Furthermore, by controlled rotation of the disc, precise and automated

fluid handling procedures are obtained. Additionally, considering its geometrical


58

symmetry, the platform is suitable for multiplexing and performing simultaneous

microfluidic processes9.

In this chapter, the sample handling-treatment element of a SERS based sensing

platform, centrifugal microfluidics will be presented.

5.1 Inertial Forces

The fluid transport and handling operations in centrifugal microfluidics are

prominently facilitated through pseudo inertial (or fictitious) forces. The fictitious forces

are characterizing a non-inertial frame of reference for which the classical Newton laws

are inapplicable. Thus, a rotating (or for broader definition with an angular acceleration)

disc which is a non-inertial frame of reference, will be influenced by these forces8,10. To

illustrate this, a specific idealized case of a rotating disc (spinning about its center) with

angular frequency 𝝎ሬሬሬԦ(𝒕) and point like mass 𝒅𝒎 of e.g. fluid on the disc with velocity 𝑽ሬሬԦ

as shown in Figure 5.1a will be taken into account. Further, interaction of the point mass

with the disc and surrounding medium as well as the fluid properties e.g. viscosity will

be neglected. In this way, the resultant inertial force on the point mass will be:

where 𝒓ሬԦ is position vector of the point mass, the first term in general bracket stands for

Coriolis, second for centrifugal and last one for the Euler forces (Figure 5.1b). In

principle, considering that the practical liquid flow rates in microfluidics systems are

generally insignificant as compared to angular frequency and acceleration of the disc, the

Coriolis force component can be neglected7. Thus, in a straightforward way, by tuning

𝑑𝐹Ԧ𝑖𝑛𝑒𝑟𝑡𝑖𝑎𝑙 = 𝑑𝑚 (2[ 𝑉ሬԦ × 𝜔ሬሬԦ ] + [ 𝜔ሬሬԦ × [ 𝑟Ԧ × 𝜔ሬሬԦ ]] + [𝑟Ԧ × (𝑑 𝜔ሬሬԦ

𝑑𝑡 )]) , (5.1)


59

the spinning parameters of the microfluidic disc platform and consequently the

centrifugal and Euler forces, desired fluid operations can be performed.

An important parameter for microfluidics based system is the driving pressure

on the fluids. In conventional devices, the precise dynamic liquid transport is realized

through controlled pressure difference acquired via a pumping equipment. Similarly, in

enclosed centrifugal microfluidics platform, the external pressure is applied through the

inertial forces. The previous case can serve as a representative example to demonstrate

this. More specifically, considering a tiny microfluidic channel filled with fluid (from 𝒓𝒊

to 𝒓𝒇) as shown in Figure 5.1a, the pressure difference between the two ends of liquid

level would be:

Figure 5.1 Pseudo forces in centrifugal microfluidics platform. (a) A disc, rotating about its

center with angular frequency 𝝎ሬሬሬԦ and with a tiny channel from 𝒓𝒊 to 𝒓𝒇 in radial direction filled

with fluid. (b) Inertial forces on a point like mass 𝒅𝒎 stimulated through the rotational motion

of the disc.


60

where 𝒅𝑽 term represents differential volume of the point like mass 𝒅𝒎 . Next, for

simplicity the rotation of the disc can be assumed to be constant which implies that the

angular acceleration would be zero. Therefore, the Euler term in Equation (5.1) will not

be considered. In addition to that, with respect to previous assumption with negligible

Coriolis force component, the pressure difference in Equation (5.2) caused by centrifugal

force can be given in a following way:

Here, 𝝆 represents the density of the fluid. The Equation (5.3) is useful when estimating

an equilibrium condition of fluid arrangement at constant rotational frequency. Hence,

the essential primary optimizations of microfluidic structures can be achieved using this

analytical formulation.

The dynamic fluid handling operations are demanding specificity in force

direction to guide the sample through the microfluidic structures in a required way. For

this reason, as an example, the Euler force component is employed to regulate the force

action on fluids in angular direction. Further, auxiliary fluid manipulations in a centrifugal

platform are accomplished through fluid-surrounding medium interactions. In the

following sections, some basic examples for those interactions will be provided.

∆𝑃 = ∫(𝑑𝐹Ԧ𝑖𝑛𝑒𝑟𝑡𝑖𝑎𝑙 ∙ 𝑑𝑟Ԧ )

𝑑𝑉

𝑟𝑓

𝑟𝑖

, (5.2)

∆𝑃 =𝜌𝜔2

2(𝑟𝑓

2 − 𝑟𝑖 2) . (5.3)


61

5.2 Pneumatic Pumping

In Section 5.1 the inertial forces acting on fluids in a rotating/accelerating disc

platform were presented. It should be emphasized that each force component in Equation

(5.1) can be altered to govern the liquid. However, by employing the pseudo forces, liquid

flow towards the center of the disc or in other words, opposite to the radial direction

cannot be realized11. This suggests that the subsequent liquid configurations in

microfluidic structures, accomplished by means of Euler and centrifugal forces, are not

reversible. On that account, various supplementary techniques are utilized together with

inertial forces in order to obtain higher degree of freedom of fluid management in a

Figure 5.2 Pneumatic pumping principle in centrifugal microfluidics. (a) A simplified

enclosed microfluidics system which consists of a sealed chamber (left) and a tiny channel with

an air vent which is linked to it (right). The stationary liquid arrangement at the rotational

frequency 𝝎𝟎 is shown. (b) Similarly, the equilibrium condition for the fluid distribution at a

higher spinning frequency 𝝎𝟏 is depicted. As a result, the liquid level and air pressure in the

sealed pneumatic chamber are increased. The associated geometrical and thermodynamical

parameters are labelled in the figure.


62

centrifugal platform. One commonly exploited method is the pneumatic pumping which

is fundamentally accomplished through compression (or decompression) of the air in an

enclosed microfluidic chamber. Thus, by manipulating the air pressure in a pneumatic

chamber, additional dynamic action on fluid samples can be obtained11–13. To interpret

the basic working principle of the technique, the microfluidic model presented in Figure

5.2 will be given as an example. Here, a straightforward microfluidic design containing a

sealed chamber (pneumatic chamber) interconnected with an air vent though a tiny

channel is shown. Further, a hypothetical liquid configuration for a specific rotational

frequency 𝝎𝟎 of the disc at the equilibrium condition is depicted in Figure 5.2a. The

stationary arrangement is summarized through thermodynamic properties of the

compressed air in the pneumatic chamber (pressure 𝑷𝟎 and corresponding volume 𝑽𝟎)

and the radial height parameters ( ∆𝒓𝟎 and 𝒓𝟎 ) of the homogeneous liquid in the

microfluidic structures. With respect to the Equation (5.3), the analytical form of this

steady state, as well as the relationship between liquid configuration and the

thermodynamical properties of the pneumatic chamber can be formulated in the following

way:

Eventually, this suggests that the action of the compressed air in pneumatic chamber

(𝑷𝟎 > 𝑷𝒂𝒕𝒎) is compensated through liquid height difference at both ends of the fluid

(centrifugal pressure, see Section 5.1). As a remark, the capillary action which occurs in

the narrow microfluidic channels was not included in the model. The effect will be

elaborated in the following Section 5.3.

∆𝑃 = 𝑃0 − 𝑃𝑎𝑡𝑚 =𝜌𝜔0

2

2 ∆𝑟0(2𝑟0 − ∆𝑟0) . (5.4)


63

Certainly, by applying additional centrifugal pressure (e.g. by increasing the

angular frequency of the disc to 𝝎𝟏 and keeping the liquid volume constant) the

thermodynamical properties of compressed air and the liquid configuration would

change. Thus, considering a quasi-static acceleration of the disc, in a similar manner the

equilibrium state at angular frequency 𝝎𝟏 can be given as (Figure 5.2b):

Since the acceleration process of the disc was assumed to be quasi-static (isothermal

procedure), the dependence of thermodynamical parameters between two distinct

stationary conditions is described through the Boyle-Mariotte’s law14:

Moreover, this thermodynamic process is theoretically reversible (due to dissipative

forces, practically it is not fully reversible) which suggests that by slowly altering the

rotational frequency of the disc the liquid heights can be precisely regulated15.

5.3 Capillary Valving

A complementary effect which has sufficient impact on microfluidic handling

procedures is the capillary action of liquids in narrow spaces (e.g. a tiny channel). This

surface phenomenon is highly dependent on fluid-environment interactions. In this way,

under influence of adhesive and cohesive forces between liquid and e.g. microfluidic

channel wells, the internal pressure value can be varied16–18. For instance, if the fluid

possesses hydrophilic properties towards the surface of the microfluidic channel

(dominating adhesive forces), the internal fluid pressure value is lowered. In contrast to

that, hydrophobic interactions would imply increase in internal pressure of the liquid.

∆𝑃 = 𝑃1 − 𝑃𝑎𝑡𝑚 =𝜌𝜔1

2

2 ∆𝑟1(2𝑟1 − ∆𝑟1) . (5.5)

𝑃0𝑉0 = 𝑃1𝑉1 . (5.6)


64

Considering the microfluidic design summarized in Figure 5.3 and relatively small

channel width 𝒅, the capillary rise or capillary pressure can be analytically defined using

Young-Laplace equality19:

where 𝜶 is the surface tension of the fluid and 𝜽 is the contact angle at the fluid-solid

interface. Subsequently, by regulating the centrifugal pressure in Equation (5.7), the

liquid level can be adjusted. The effect of capillary rise is frequently utilized in a

centrifugal microfluidic platform for valving techniques. Thus, a carefully designed

microfluidic channel (geometrical shape, dimensions, material and possible coating) can

∆𝑃 =𝜌𝜔2

2 ∆𝑟(2𝑟 − ∆𝑟) =

4𝛼 𝑐𝑜𝑠𝜃

𝑑 , (5.7)

Figure 5.3 Capillary action in centrifugal microfluidics. The fluid arrangement at the

equilibrium condition for rotational angular frequency 𝝎. Due to air vent openings, the pressure

values at both liquid ends are constant. The capillary interaction (in this case hydrophilic)

facilitates rise of fluid height in a narrow channel which has a smaller width than the chamber.


65

serve as a conditional (rotational parameters) fluid transfer unit e.g. from one chamber to

another8,11.

5.4 Design and Fabrication of Microfluidic Discs

In an accurate way, microfluidic structures for specific sample handling

operations are designed using computer modelling techniques. Accordingly, by carrying

out computational simulations of fluid interactions with surrounding media, an estimated

fluid behavior and consequently an optimization process of the microfluidic structure

geometry is performed. Following that, using fabricated prototype discs, additional and

final adjustments in structure morphology are made20,21. Nevertheless, the computational

techniques, specifically complete 3-dimensional fluidic simulations are time consuming

and it is often the case that the results are depicting inaccurate microfluidic response (due

to errors in computational parameters).

Throughout this PhD study, the design procedure for microfluidic structures was

conducted in an alternative way. First of all, using the analytical assumptions described

in Sections 5.1, 5.2 and 5.3, a rough approximation of a microfluidic model for desired

handling operations was obtained. Further, by taking advantage of the rapid prototyping

fabrication technique, described in Chapter 0, the optimization and fine adjustment of

microfluidic designs were achieved experimentally. The discs were produced via step-

by-step assembly of laser cut polymeric discs (e.g. poly (methyl methacrylate)-PMMA)

and patterned pressure sensitive adhesive (PSA) layers. The complete SERS integrated

centrifugal microfluidic discs are produced within 20-30 min. Lastly, the desired fluid

handling operations are tested in an optical spin stand where real-time images of the disc

are captured during the centrifugation process. After each design and test, necessary

adjustments for desired fluidic operations were implemented for the next prototypes.


66

Finally, the testified and optimized design can be produced in a facile and robust

way using mass production methods. Injection molding as an exemplary mass production

technique is demonstrated in Chapter 0.

References

(1) Whitesides, G. M. Nature 2006, 442 (7101), 368–373.

(2) Squires, T. M.; Quake, S. R. Rev. Mod. Phys. 2005, 77 (3), 977–1026.

(3) Stone, H. A.; Stroock, A. D.; Ajdari, A. Annu. Rev. Fluid Mech. 2004, 36 (1),

381–411.

(4) Beebe, D. J.; Mensing, G. A.; Walker, G. M. Annu. Rev. Biomed. Eng. 2002, 4

(1), 261–286.

(5) Wang, G.; Ho, H.-P.; Chen, Q.; Yang, A. K.-L.; Kwok, H.-C.; Wu, S.-Y.; Kong,

S.-K.; Kwan, Y.-W.; Zhang, X. Lab Chip 2013, 13 (18), 3698.

(6) Strohmeier, O.; Keller, M.; Schwemmer, F.; Zehnle, S.; Mark, D.; von Stetten,

F.; Zengerle, R.; Paust, N. Chem. Soc. Rev. 2015, 44 (17), 6187–6229.

(7) Gorkin, R.; Park, J.; Siegrist, J.; Amasia, M.; Lee, B. S.; Park, J.-M.; Kim, J.;

Kim, H.; Madou, M.; Cho, Y.-K. Lab Chip 2010, 10 (14), 1758.

(8) Ducrée, J.; Haeberle, S.; Lutz, S.; Pausch, S.; Stetten, F. von; Zengerle, R. J.

Micromechanics Microengineering 2007, 17 (7), S103–S115.

(9) Nolte, D. D. Rev. Sci. Instrum. 2009, 80 (10), 101101.

(10) Burger, R.; Ducrée, J. Expert Rev. Mol. Diagn. 2012, 12 (4), 407–421.

(11) Gorkin, R.; Clime, L.; Madou, M.; Kido, H. Microfluid. Nanofluidics 2010, 9

(2–3), 541–549.

(12) Mark, D.; Metz, T.; Haeberle, S.; Lutz, S.; Ducrée, J.; Zengerle, R.; von Stetten,

F. Lab Chip 2009, 9 (24), 3599.

(13) Abi-Samra, K.; Clime, L.; Kong, L.; Gorkin, R.; Kim, T.-H.; Cho, Y.-K.; Madou,

M. Microfluid. Nanofluidics 2011, 11 (5), 643–652.

(14) De Groot, S. R.; Mazur, P. Non-equilibrium thermodynamics; Courier

Corporation, 2013.


67

(15) Gorkin, R.; Nwankire, C. E.; Gaughran, J.; Zhang, X.; Donohoe, G. G.; Rook,

M.; O’Kennedy, R.; Ducrée, J. Lab Chip 2012, 12 (16), 2894.

(16) Chen, J. M.; Huang, P.-C.; Lin, M.-G. Microfluid. Nanofluidics 2008, 4 (5), 427–

437.

(17) Siegrist, J.; Gorkin, R.; Clime, L.; Roy, E.; Peytavi, R.; Kido, H.; Bergeron, M.;

Veres, T.; Madou, M. Microfluid. Nanofluidics 2010, 9 (1), 55–63.

(18) Zeng, J.; Banerjee, D.; Deshpande, M.; Gilbert, J. R.; Duffy, D. C.; Kellogg, G.

J. In Micro Total Analysis Systems 2000; Springer Netherlands: Dordrecht, 2000; pp 579–

582.

(19) Cho, H.; Kim, H.-Y.; Kang, J. Y.; Kim, T. S. J. Colloid Interface Sci. 2007, 306

(2), 379–385.

(20) Rasmussen, A.; Mavriplis, C.; Zaghloul, M. E.; Mikulchenko, O.; Mayaram, K.

Sensors Actuators A Phys. 2001, 88 (2), 121–132.

(21) Chen, X.; Cui, D.; Liu, C.; Li, H.; Chen, J. Anal. Chim. Acta 2007, 584 (2), 237–

243.


68

Chapter 6: Nanopillar Filters for SERS

69

6 NANOPILLAR FILTERS FOR SERS

Original title: Nanopillar filters for Surface-Enhanced Raman Spectroscopy

Onur Durucan1, *, Tomas Rindzevicius1, Michael Stenbæk Schmidt1, Marco Matteucci1, and Anja Boisen1

1Technical University of Denmark, Department of Micro-and Nano Technology, Kgs. Lyngby, 2800, Denmark

*[email protected]

Published in ACS Sens., 2017, 2 (10), pp 1400–1404.

Abstract

We present a simple, robust and automated molecule extraction technique based on a

centrifugal microfluidic platform. Fast and facile extraction of a food adulterant (melamine) from

a complex sample medium (milk) on a SERS substrate is demonstrated. The unique characteristic

of the detection method is the obtained “filter paper/chromatographic” effect which combines

centrifugal force and wetting properties of the SERS substrate. The work addresses issues related

to SERS-based detection of analytes in complex media, which is important for realizing next

generation SERS platforms applicable for a broad variety of real-life applications.

Keywords: SERS, Wicking, Microfluidics, Nanopillars, Filtration, Melamine

Availablehttps://pubs.acs.org/doi/abs/10.1021/acssensors.7b00499

online:

Chapter 7: NP Structures for SERS Chromatography

91

7 NP STRUCTURES FOR SERS

CHROMATOGRAPHY

Original title: Nanopillar structures for Surface-Enhanced Raman Spectroscopy

Chromatography

Onur Durucan1, *, Kaiyu Wu1, *, Tomas Rindzevicius1, Marlitt Viehrig1, Oleksii Ilchenko1, Michael Stenbæk

Schmidt1, and Anja Boisen1


*[email protected], *[email protected]

Unpublished Manuscript.

Abstract

We demonstrate a new, fully integrated chromatography method for analyzing complex,

multicomponent samples, i.e. human urine spiked with paracetamol, using surface-enhanced

Raman spectroscopy (SERS). Gold coated silicon nanopillar substrates (Au NP) and disc-based,

centrifugal microfluidic platform are combined to manipulate the liquid matrix and

simultaneously generate a chromatographic effect on the ~4x4 mm Au NP surface area. Results

show that high density (~48 pillars/ µm-2) Au NP structures can be used to effectively separate

paracetamol and the main human urine components (urea, uric acid and creatinine) on the

substrate for subsequent SERS detection. The Au NP induced chromatographic effect was

visualized and analyzed using the multivariate curve resolution-alternative least squares (MCR-

ALS) method. The integrated SERS chromatography platform was further utilized for quantitative

detection of paracetamol (0-500 ppm) in human urine. By analyzing the analyte surface spreading

profiles, a reproducible, linear trend across the entire concentration range was obtained. The new

method is an important leap towards developing a fully integrated SERS sensing platform for

quantitative, multicomponent analysis of analytes in various biofluids e.g. saliva, blood or urine.

Keywords: SERS, Wicking, Microfluidics, Nanopillars, Chromatography, Urine, Paracetamol

Unpublished Manuscript

Chapter 8: Dual-Functional EC and SERS Sensing

121

8 DUAL-FUNCTIONAL EC AND SERS

SENSING

Original title: Large-scale, Lithography-free Production of Transparent Nanostructured Surface

for Dual-functional Electrochemical and SERS Sensing

Kuldeep Sanger1, *, Onur Durucan1, *, Kaiyu Wu1, Anil Haraksingh Thilsted1, Arto Heiskanen1, Tomas

Rindzevicius1, Michael Stenbæk Schmidt1, Kinga Zor1 and Anja Boisen1


*[email protected], *[email protected]

Published in ACS Sens., 2017, 2 (12), pp 1869–1875

Abstract

In this work, we present a dual-functional sensor that can perform surface-enhanced

Raman spectroscopy (SERS) based identification and electrochemical (EC) quantification of

analytes in liquid samples. A lithography-free reactive ion etching process was utilized to obtain

nanostructures of high aspect ratios distributed homogeneously on a 4-inch fused silica wafer.

The sensor was made up of three-electrode array, obtained by subsequent e-beam evaporation of

Au on nanostructures in selected areas through a shadow mask. The SERS performance was

evaluated through surface-averaged enhancement factor (EF), which was ~6.2 x 105, and spatial

uniformity of EF, which was ~13% in terms of relative standard deviation. Excellent

electrochemical performance and reproducibility were revealed by recording cyclic

voltammograms. On nanostructured electrodes, paracetamol (PAR) showed an improved quasi-

reversible behavior with decrease in peak potential separation (∆Ep ~90mV) and higher peak

currents (Ipa/Ipc ~1), comparing to planar electrodes (∆Ep ~560mV). The oxidation potential of

PAR was also lowered by ~80 mV on nanostructured electrodes. To illustrate dual-functional

sensing, quantitative evaluation of PAR ranging from 30 µM to 3 mM was realized through EC

detection, and presence of PAR was verified by its SERS fingerprint.

Keywords: Lithography-free, dual-functional, electrochemical, SERS, paracetamol

Availablehttps://pubs.acs.org/doi/abs/10.1021/acssensors.7b00783

online:

Chapter 9: Concluding Remarks

141

9 CONCLUDING REMARKS

In this PhD thesis, applicability of surface-enhanced Raman spectroscopy

(SERS) based centrifugal microfluidic platform for sensing applications was investigated.

The aim of the platform is to automate sample pre-treatment (e.g. filtration, separation,

etc.) procedure and enable quantitative SERS-based analysis of molecular species in

chemically rich, complex, real-life samples. As discussed in Chapter 4, it is well-known

that this requires (1) reliable and high-performance SERS substrates, (2) advanced sample

pre-treatment and dosing techniques. With the recent developments in nanofabrication

methods, the first point can be addressed by utilizing highly uniform nanopillar (NP)

structures fabricated on large-scale surfaces (4 inch wafers). The gold metal coated NP

arrays developed in the Nanoprobes research group exhibit exceptionally good SERS

performance. Thus, the focus for the PhD study was on integration of the NP SERS

substrate into a microfluidics platform and application development.

The sample handling and pre-treatment steps were accomplished in the SERS

integrated centrifugal microfluidics platform (see Chapter 6 and 7). By taking advantage

of precise and automated fluid management procedures in microfluidics platform,

reproducible incubation of sample onto the NP surface can be obtained. On that account,

various microfluidic techniques such as metering, valving and pneumatic pumping were

realized. Another crucial impact of the platform is the regulative inertial forces which are

facilitated through rotation of the microfluidic discs. By manipulating the rotational

parameters and consequently the inertial forces, an optional sedimentation of

macromolecular particles in sample suspension (centrifugation) was carried out. In an


142

illustrative way, this was demonstrated on melamine doped milk solutions under 47.5 Hz

rotational frequency of the disc.

The centrifugation process promoted phase separation of the sample with respect

to density gradient of the contents. However, the centrifugation as pre-treatment method

was not sufficient enough to obtain the required degree of sample purity. Considering the

high sensitivity and selectivity of the surface-enhanced Raman scattering (RS)

phenomenon, the real-life applications require (i) complete filtration of macromolecules,

(ii) separation of distinct analyte species and lastly (iii) uniform analyte dosage on the

surface. On this account, the wicking effect caused by capillary interaction of fluid sample

and nanotextured surface (NP structures) was employed in the centrifugal platform.

Results in the first study showed that NP structures can be used as nanofilters

for pre-treating milk samples which significantly improves melamine detection using

SERS (Chapter 6). Macromolecular clusters such as proteins and fats were clogged in

the immersion region while the diffusing thin layer of the milk suspension (through

capillary wicking) was purified. This was verified using morphological inspection of the

NP surface through scanning electron microscopy. Furthermore, the effect facilitated

controlled deposition of analyte molecules (artificially doped melamine) onto the NP

SERS substrate. The results from experimental SERS mapping on purified region of NP

surface indicated that the melamine molecules were uniformly accumulated. Following

that, the statistical SERS map studies of analyte peak at 687 cm-1 demonstrated that the

quantification of melamine concentrations ranging from 10 parts per million (ppm) up to

400 ppm is possible using this technique. Lastly, melamine concentration versus the

SERS signal intensity response obtained using the novel sample dosing technique was in

agreement with Langmuir adsorption curve which is a characteristic feature of the SERS

based sensing platforms.


143

In the second case study, the aim was to obtain a “chromatographic effect” from

paracetamol doped human urine samples directly on the nanopillar SERS substrate for

simultaneous SERS detection of the analyte and different urine compounds. Essentially,

the main challenge for a reliable SERS based sensing of analytes in a complex matrix

arises from the competitive binding effect of sample contents on the SERS active surface.

The surface-enhanced RS phenomenon is highly surface sensitive, thus only the

molecules which are located in the close vicinity of the active region (electromagnetic

“hot-spots”) will contribute to the SERS signal. The distribution of molecular compounds

in complex sample suspension on SERS active region is highly determined by their

affinities towards the plasmonic surface. Practically, the application of multicomponent

sample solution on a SERS substrate without any sample pre-treatment results in selective

accumulation of components (molecule species with higher affinity) on SERS active

structures. For the “chromatographic effect” described in Chapter 7, wicking

phenomenon which proposes gradual diffusion of fluid sample on the NP surface was

utilized to segregate molecular species in the multicomponent sample solution. More

specifically, along the spreading path of urine solution on the NP substrate, the

arrangement of components was determined by their binding property (affinity) towards

plasmonic surface (Au). Consequently, the contents were trapped locally on SERS active

regions. Further, the experimental SERS studies and the spectral decomposition analysis

confirmed that the typical components of human urine such as urea, uric acid, creatinine

and the artificially doped paracetamol were successfully separated along the NP SERS

substrate. Moreover, a linear concentration dependence of paracetamol with respect to its

spectral profile was shown. Accordingly, quantification of paracetamol concentration

ranging from 30 to 500 ppm was demonstrated.

Finally, an alternative dual functional sensing approach was experimentally

investigated (Chapter 8). The capability of electrochemical (EC) sensors to quantify


144

analyte molecules through reduction-oxidation reactions and spectral specificity of the

SERS technique was combined into a single detection platform. The goal was to explore

if the two techniques could complement each other and improve the analyte detection in

liquid samples. In order to integrate the two techniques into a single sensing unit, NP-like

structures on fused silica suitable for both EC and SERS detection were fabricated and

utilized as an active surface. The advantages and limitations of the proposed dual-

functioning sensor concept was examined using phosphate-buffered saline (PBS) solution

doped with paracetamol. The EC quantification and SERS identification limit was

estimated to be 30 µM paracetamol concentration in PBS.

In summary, the thesis presents a number of solutions to address commonly

encountered issues associated with SERS based sensors, especially when dealing with

real-life samples such as milk, or complex biofluids such as urine. The research conducted

during the PhD project is aimed at the development of next generation SERS platforms

with automated sample manipulation, pre-treatment and filtration. The suggested

methodologies are suitable for a wide range of practical application, and could be further

optimized by e.g. tuning the morphology of SERS active structures. Importantly, results

showed that the SERS-centrifugal microfluidic platform can be used for multicomponent,

label-free and sensitive detection of analyte species in a complex matrix without using

any surface functionalization procedures.


145


146

Surface-Enhanced Raman Spectroscopy Integrated Centrifugal Microfluidics Platform ·...

Documents

Transcript of Surface-Enhanced Raman Spectroscopy Integrated Centrifugal Microfluidics Platform ·...